{ "cells": [ { "cell_type": "markdown", "id": "e4d57d53", "metadata": {}, "source": [ "# Divoce" ] }, { "cell_type": "code", "execution_count": null, "id": "9G1AOK-PWzwR", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 25114, "status": "ok", "timestamp": 1746448292946, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "9G1AOK-PWzwR", "outputId": "eebdb5f0-4e07-4e23-edbd-0bf27f883f60" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Requirement already satisfied: jieba in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (0.42.1)\n", "Requirement already satisfied: gensim==4.3.3 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (4.3.3)\n", "Collecting spacy==3.7.2\n", " Downloading spacy-3.7.2-cp311-cp311-win_amd64.whl.metadata (26 kB)\n", "Collecting thinc==8.2.2\n", " Downloading thinc-8.2.2-cp311-cp311-win_amd64.whl.metadata (15 kB)\n", "Requirement already satisfied: numpy<2.0,>=1.18.5 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from gensim==4.3.3) (1.24.3)\n", "Requirement already satisfied: scipy<1.14.0,>=1.7.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from gensim==4.3.3) (1.12.0)\n", "Requirement already satisfied: smart-open>=1.8.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from gensim==4.3.3) (7.1.0)\n", "Collecting spacy-legacy<3.1.0,>=3.0.11 (from spacy==3.7.2)\n", " Downloading spacy_legacy-3.0.12-py2.py3-none-any.whl.metadata (2.8 kB)\n", "Collecting spacy-loggers<2.0.0,>=1.0.0 (from spacy==3.7.2)\n", " Downloading spacy_loggers-1.0.5-py3-none-any.whl.metadata (23 kB)\n", "Collecting murmurhash<1.1.0,>=0.28.0 (from spacy==3.7.2)\n", " Downloading murmurhash-1.0.12-cp311-cp311-win_amd64.whl.metadata (2.2 kB)\n", "Collecting cymem<2.1.0,>=2.0.2 (from spacy==3.7.2)\n", " Downloading cymem-2.0.11-cp311-cp311-win_amd64.whl.metadata (8.8 kB)\n", "Collecting preshed<3.1.0,>=3.0.2 (from spacy==3.7.2)\n", " Downloading preshed-3.0.9-cp311-cp311-win_amd64.whl.metadata (2.2 kB)\n", "Collecting wasabi<1.2.0,>=0.9.1 (from spacy==3.7.2)\n", " Downloading wasabi-1.1.3-py3-none-any.whl.metadata (28 kB)\n", "Collecting srsly<3.0.0,>=2.4.3 (from spacy==3.7.2)\n", " Downloading srsly-2.5.1-cp311-cp311-win_amd64.whl.metadata (20 kB)\n", "Collecting catalogue<2.1.0,>=2.0.6 (from spacy==3.7.2)\n", " Downloading catalogue-2.0.10-py3-none-any.whl.metadata (14 kB)\n", "Collecting weasel<0.4.0,>=0.1.0 (from spacy==3.7.2)\n", " Downloading weasel-0.3.4-py3-none-any.whl.metadata (4.7 kB)\n", "Collecting typer<0.10.0,>=0.3.0 (from spacy==3.7.2)\n", " Downloading typer-0.9.4-py3-none-any.whl.metadata (14 kB)\n", "Collecting smart-open>=1.8.1 (from gensim==4.3.3)\n", " Downloading smart_open-6.4.0-py3-none-any.whl.metadata (21 kB)\n", "Requirement already satisfied: tqdm<5.0.0,>=4.38.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from spacy==3.7.2) (4.67.1)\n", "Requirement already satisfied: requests<3.0.0,>=2.13.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from spacy==3.7.2) (2.32.3)\n", "Collecting pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4 (from spacy==3.7.2)\n", " Downloading pydantic-2.11.4-py3-none-any.whl.metadata (66 kB)\n", "Requirement already satisfied: jinja2 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from spacy==3.7.2) (3.1.6)\n", "Requirement already satisfied: setuptools in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from spacy==3.7.2) (75.8.0)\n", "Requirement already satisfied: packaging>=20.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from spacy==3.7.2) (24.2)\n", "Collecting langcodes<4.0.0,>=3.2.0 (from spacy==3.7.2)\n", " Downloading langcodes-3.5.0-py3-none-any.whl.metadata (29 kB)\n", "Collecting blis<0.8.0,>=0.7.8 (from thinc==8.2.2)\n", " Downloading blis-0.7.11-cp311-cp311-win_amd64.whl.metadata (7.6 kB)\n", "Collecting confection<1.0.0,>=0.0.1 (from thinc==8.2.2)\n", " Downloading confection-0.1.5-py3-none-any.whl.metadata (19 kB)\n", "Collecting language-data>=1.2 (from langcodes<4.0.0,>=3.2.0->spacy==3.7.2)\n", " Downloading language_data-1.3.0-py3-none-any.whl.metadata (4.3 kB)\n", "Collecting annotated-types>=0.6.0 (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy==3.7.2)\n", " Downloading annotated_types-0.7.0-py3-none-any.whl.metadata (15 kB)\n", "Collecting pydantic-core==2.33.2 (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy==3.7.2)\n", " Downloading pydantic_core-2.33.2-cp311-cp311-win_amd64.whl.metadata (6.9 kB)\n", "Requirement already satisfied: typing-extensions>=4.12.2 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy==3.7.2) (4.12.2)\n", "Collecting typing-inspection>=0.4.0 (from pydantic!=1.8,!=1.8.1,<3.0.0,>=1.7.4->spacy==3.7.2)\n", " Downloading typing_inspection-0.4.0-py3-none-any.whl.metadata (2.6 kB)\n", "Requirement already satisfied: charset-normalizer<4,>=2 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests<3.0.0,>=2.13.0->spacy==3.7.2) (3.3.2)\n", "Requirement already satisfied: idna<4,>=2.5 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests<3.0.0,>=2.13.0->spacy==3.7.2) (3.7)\n", "Requirement already satisfied: urllib3<3,>=1.21.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests<3.0.0,>=2.13.0->spacy==3.7.2) (2.3.0)\n", "Requirement already satisfied: certifi>=2017.4.17 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests<3.0.0,>=2.13.0->spacy==3.7.2) (2025.1.31)\n", "Requirement already satisfied: colorama in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from tqdm<5.0.0,>=4.38.0->spacy==3.7.2) (0.4.6)\n", "Requirement already satisfied: click<9.0.0,>=7.1.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from typer<0.10.0,>=0.3.0->spacy==3.7.2) (8.1.8)\n", "Collecting cloudpathlib<0.17.0,>=0.7.0 (from weasel<0.4.0,>=0.1.0->spacy==3.7.2)\n", " Downloading cloudpathlib-0.16.0-py3-none-any.whl.metadata (14 kB)\n", "Requirement already satisfied: MarkupSafe>=2.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from jinja2->spacy==3.7.2) (3.0.2)\n", "Collecting marisa-trie>=1.1.0 (from language-data>=1.2->langcodes<4.0.0,>=3.2.0->spacy==3.7.2)\n", " Downloading marisa_trie-1.2.1-cp311-cp311-win_amd64.whl.metadata (9.3 kB)\n", "Downloading spacy-3.7.2-cp311-cp311-win_amd64.whl (12.1 MB)\n", " ---------------------------------------- 0.0/12.1 MB ? eta -:--:--\n", " --------------------------------------- 0.3/12.1 MB ? eta -:--:--\n", " - -------------------------------------- 0.5/12.1 MB 1.3 MB/s eta 0:00:09\n", " --- ------------------------------------ 1.0/12.1 MB 1.9 MB/s eta 0:00:06\n", " ----- ---------------------------------- 1.6/12.1 MB 2.2 MB/s eta 0:00:05\n", " ------ --------------------------------- 2.1/12.1 MB 2.2 MB/s eta 0:00:05\n", " -------- ------------------------------- 2.6/12.1 MB 2.3 MB/s eta 0:00:05\n", " --------- ------------------------------ 2.9/12.1 MB 2.3 MB/s eta 0:00:04\n", " ------------ --------------------------- 3.7/12.1 MB 2.3 MB/s eta 0:00:04\n", " ------------- -------------------------- 3.9/12.1 MB 2.3 MB/s eta 0:00:04\n", " -------------- ------------------------- 4.5/12.1 MB 2.3 MB/s eta 0:00:04\n", " ---------------- ----------------------- 5.0/12.1 MB 2.2 MB/s eta 0:00:04\n", " ----------------- ---------------------- 5.2/12.1 MB 2.2 MB/s eta 0:00:04\n", " ------------------- -------------------- 5.8/12.1 MB 2.2 MB/s eta 0:00:03\n", " -------------------- ------------------- 6.3/12.1 MB 2.2 MB/s eta 0:00:03\n", " --------------------- ------------------ 6.6/12.1 MB 2.2 MB/s eta 0:00:03\n", " ---------------------- ----------------- 6.8/12.1 MB 2.1 MB/s eta 0:00:03\n", " ----------------------- ---------------- 7.1/12.1 MB 2.0 MB/s eta 0:00:03\n", " ------------------------ --------------- 7.3/12.1 MB 2.0 MB/s eta 0:00:03\n", " ------------------------- -------------- 7.6/12.1 MB 2.0 MB/s eta 0:00:03\n", " -------------------------- ------------- 8.1/12.1 MB 2.0 MB/s eta 0:00:03\n", " ---------------------------- ----------- 8.7/12.1 MB 2.0 MB/s eta 0:00:02\n", " ------------------------------ --------- 9.2/12.1 MB 2.0 MB/s eta 0:00:02\n", " ------------------------------- -------- 9.4/12.1 MB 2.0 MB/s eta 0:00:02\n", " --------------------------------- ------ 10.0/12.1 MB 2.0 MB/s eta 0:00:02\n", " ---------------------------------- ----- 10.5/12.1 MB 2.0 MB/s eta 0:00:01\n", " ------------------------------------ --- 11.0/12.1 MB 2.0 MB/s eta 0:00:01\n", " -------------------------------------- - 11.5/12.1 MB 2.0 MB/s eta 0:00:01\n", " ---------------------------------------- 12.1/12.1 MB 2.0 MB/s eta 0:00:00\n", "Downloading thinc-8.2.2-cp311-cp311-win_amd64.whl (1.5 MB)\n", " ---------------------------------------- 0.0/1.5 MB ? eta -:--:--\n", " ------- -------------------------------- 0.3/1.5 MB ? eta -:--:--\n", " ------- -------------------------------- 0.3/1.5 MB ? eta -:--:--\n", " ----------------------------------- ---- 1.3/1.5 MB 2.2 MB/s eta 0:00:01\n", " ---------------------------------------- 1.5/1.5 MB 1.9 MB/s eta 0:00:00\n", "Downloading blis-0.7.11-cp311-cp311-win_amd64.whl (6.6 MB)\n", " ---------------------------------------- 0.0/6.6 MB ? eta -:--:--\n", " - -------------------------------------- 0.3/6.6 MB ? eta -:--:--\n", " ------ --------------------------------- 1.0/6.6 MB 2.6 MB/s eta 0:00:03\n", " ------- -------------------------------- 1.3/6.6 MB 2.1 MB/s eta 0:00:03\n", " --------- ------------------------------ 1.6/6.6 MB 2.2 MB/s eta 0:00:03\n", " ------------ --------------------------- 2.1/6.6 MB 2.1 MB/s eta 0:00:03\n", " -------------- ------------------------- 2.4/6.6 MB 2.0 MB/s eta 0:00:03\n", " ------------------- -------------------- 3.1/6.6 MB 2.2 MB/s eta 0:00:02\n", " ---------------------- ----------------- 3.7/6.6 MB 2.2 MB/s eta 0:00:02\n", " ------------------------- -------------- 4.2/6.6 MB 2.3 MB/s eta 0:00:02\n", " -------------------------- ------------- 4.5/6.6 MB 2.3 MB/s eta 0:00:01\n", " ------------------------------ --------- 5.0/6.6 MB 2.2 MB/s eta 0:00:01\n", " ------------------------------- -------- 5.2/6.6 MB 2.1 MB/s eta 0:00:01\n", " ---------------------------------- ----- 5.8/6.6 MB 2.1 MB/s eta 0:00:01\n", " --------------------------------------- 6.6/6.6 MB 2.2 MB/s eta 0:00:01\n", " ---------------------------------------- 6.6/6.6 MB 2.1 MB/s eta 0:00:00\n", "Downloading catalogue-2.0.10-py3-none-any.whl (17 kB)\n", "Downloading confection-0.1.5-py3-none-any.whl (35 kB)\n", "Downloading cymem-2.0.11-cp311-cp311-win_amd64.whl (39 kB)\n", "Downloading langcodes-3.5.0-py3-none-any.whl (182 kB)\n", "Downloading murmurhash-1.0.12-cp311-cp311-win_amd64.whl (25 kB)\n", "Downloading preshed-3.0.9-cp311-cp311-win_amd64.whl (122 kB)\n", "Downloading pydantic-2.11.4-py3-none-any.whl (443 kB)\n", "Downloading pydantic_core-2.33.2-cp311-cp311-win_amd64.whl (2.0 MB)\n", " ---------------------------------------- 0.0/2.0 MB ? eta -:--:--\n", " ---------------------------------------- 0.0/2.0 MB ? eta -:--:--\n", " ---------- ----------------------------- 0.5/2.0 MB 1.5 MB/s eta 0:00:01\n", " --------------------- ------------------ 1.0/2.0 MB 2.3 MB/s eta 0:00:01\n", " -------------------------- ------------- 1.3/2.0 MB 2.0 MB/s eta 0:00:01\n", " ------------------------------------- -- 1.8/2.0 MB 1.9 MB/s eta 0:00:01\n", " ---------------------------------------- 2.0/2.0 MB 1.7 MB/s eta 0:00:00\n", "Downloading smart_open-6.4.0-py3-none-any.whl (57 kB)\n", "Downloading spacy_legacy-3.0.12-py2.py3-none-any.whl (29 kB)\n", "Downloading spacy_loggers-1.0.5-py3-none-any.whl (22 kB)\n", "Downloading srsly-2.5.1-cp311-cp311-win_amd64.whl (632 kB)\n", " ---------------------------------------- 0.0/632.6 kB ? eta -:--:--\n", " ---------------- ----------------------- 262.1/632.6 kB ? eta -:--:--\n", " --------------------------------- ------ 524.3/632.6 kB 1.4 MB/s eta 0:00:01\n", " ---------------------------------------- 632.6/632.6 kB 1.2 MB/s eta 0:00:00\n", "Downloading typer-0.9.4-py3-none-any.whl (45 kB)\n", "Downloading wasabi-1.1.3-py3-none-any.whl (27 kB)\n", "Downloading weasel-0.3.4-py3-none-any.whl (50 kB)\n", "Downloading annotated_types-0.7.0-py3-none-any.whl (13 kB)\n", "Downloading cloudpathlib-0.16.0-py3-none-any.whl (45 kB)\n", "Downloading language_data-1.3.0-py3-none-any.whl (5.4 MB)\n", " ---------------------------------------- 0.0/5.4 MB ? eta -:--:--\n", " ---------------------------------------- 0.0/5.4 MB ? eta -:--:--\n", " - -------------------------------------- 0.3/5.4 MB ? eta -:--:--\n", " ----- ---------------------------------- 0.8/5.4 MB 1.4 MB/s eta 0:00:04\n", " ------- -------------------------------- 1.0/5.4 MB 1.6 MB/s eta 0:00:03\n", " --------- ------------------------------ 1.3/5.4 MB 1.3 MB/s eta 0:00:04\n", " --------- ------------------------------ 1.3/5.4 MB 1.3 MB/s eta 0:00:04\n", " ------------- -------------------------- 1.8/5.4 MB 1.3 MB/s eta 0:00:03\n", " --------------- ------------------------ 2.1/5.4 MB 1.4 MB/s eta 0:00:03\n", " ------------------- -------------------- 2.6/5.4 MB 1.5 MB/s eta 0:00:02\n", " ----------------------- ---------------- 3.1/5.4 MB 1.6 MB/s eta 0:00:02\n", " ------------------------- -------------- 3.4/5.4 MB 1.6 MB/s eta 0:00:02\n", " ----------------------------- ---------- 3.9/5.4 MB 1.6 MB/s eta 0:00:01\n", " ------------------------------- -------- 4.2/5.4 MB 1.6 MB/s eta 0:00:01\n", " ----------------------------------- ---- 4.7/5.4 MB 1.6 MB/s eta 0:00:01\n", " ------------------------------------ --- 5.0/5.4 MB 1.6 MB/s eta 0:00:01\n", " -------------------------------------- - 5.2/5.4 MB 1.6 MB/s eta 0:00:01\n", " ---------------------------------------- 5.4/5.4 MB 1.6 MB/s eta 0:00:00\n", "Downloading typing_inspection-0.4.0-py3-none-any.whl (14 kB)\n", "Downloading marisa_trie-1.2.1-cp311-cp311-win_amd64.whl (152 kB)\n", "Installing collected packages: cymem, wasabi, typing-inspection, spacy-loggers, spacy-legacy, smart-open, pydantic-core, murmurhash, marisa-trie, cloudpathlib, catalogue, blis, annotated-types, typer, srsly, pydantic, preshed, language-data, langcodes, confection, weasel, thinc, spacy\n", " Attempting uninstall: smart-open\n", " Found existing installation: smart-open 7.1.0\n", " Uninstalling smart-open-7.1.0:\n", " Successfully uninstalled smart-open-7.1.0\n", "Successfully installed annotated-types-0.7.0 blis-0.7.11 catalogue-2.0.10 cloudpathlib-0.16.0 confection-0.1.5 cymem-2.0.11 langcodes-3.5.0 language-data-1.3.0 marisa-trie-1.2.1 murmurhash-1.0.12 preshed-3.0.9 pydantic-2.11.4 pydantic-core-2.33.2 smart-open-6.4.0 spacy-3.7.2 spacy-legacy-3.0.12 spacy-loggers-1.0.5 srsly-2.5.1 thinc-8.2.2 typer-0.9.4 typing-inspection-0.4.0 wasabi-1.1.3 weasel-0.3.4\n" ] } ], "source": [ "#!pip install jieba\n", "#!pip install \"gensim==4.3.3\" \"spacy==3.7.2\" \"thinc==8.2.2\"" ] }, { "cell_type": "code", "execution_count": null, "id": "5bdb9c23", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Collecting plotly\n", " Downloading plotly-6.0.1-py3-none-any.whl.metadata (6.7 kB)\n", "Collecting narwhals>=1.15.1 (from plotly)\n", " Downloading narwhals-1.38.2-py3-none-any.whl.metadata (9.4 kB)\n", "Requirement already satisfied: packaging in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from plotly) (24.2)\n", "Downloading plotly-6.0.1-py3-none-any.whl (14.8 MB)\n", " ---------------------------------------- 0.0/14.8 MB ? eta -:--:--\n", " ---------------------------------------- 0.0/14.8 MB ? eta -:--:--\n", " --------------------------------------- 0.3/14.8 MB ? eta -:--:--\n", " - -------------------------------------- 0.5/14.8 MB 1.2 MB/s eta 0:00:12\n", " -- ------------------------------------- 0.8/14.8 MB 1.3 MB/s eta 0:00:11\n", " --- ------------------------------------ 1.3/14.8 MB 1.4 MB/s eta 0:00:10\n", " ---- ----------------------------------- 1.6/14.8 MB 1.4 MB/s eta 0:00:10\n", " ------ --------------------------------- 2.4/14.8 MB 1.7 MB/s eta 0:00:08\n", " ------- -------------------------------- 2.9/14.8 MB 1.8 MB/s eta 0:00:07\n", " -------- ------------------------------- 3.1/14.8 MB 1.8 MB/s eta 0:00:07\n", " --------- ------------------------------ 3.7/14.8 MB 1.8 MB/s eta 0:00:07\n", " ---------- ----------------------------- 3.9/14.8 MB 1.8 MB/s eta 0:00:06\n", " ------------ --------------------------- 4.5/14.8 MB 1.9 MB/s eta 0:00:06\n", " ------------- -------------------------- 5.0/14.8 MB 1.9 MB/s eta 0:00:06\n", " -------------- ------------------------- 5.5/14.8 MB 2.0 MB/s eta 0:00:05\n", " --------------- ------------------------ 5.8/14.8 MB 1.9 MB/s eta 0:00:05\n", " ---------------- ----------------------- 6.3/14.8 MB 2.0 MB/s eta 0:00:05\n", " ----------------- ---------------------- 6.6/14.8 MB 2.0 MB/s eta 0:00:05\n", " ------------------- -------------------- 7.1/14.8 MB 2.0 MB/s eta 0:00:04\n", " --------------------- ------------------ 7.9/14.8 MB 2.0 MB/s eta 0:00:04\n", " ---------------------- ----------------- 8.4/14.8 MB 2.1 MB/s eta 0:00:04\n", " ----------------------- ---------------- 8.7/14.8 MB 2.1 MB/s eta 0:00:03\n", " ------------------------ --------------- 9.2/14.8 MB 2.1 MB/s eta 0:00:03\n", " ------------------------- -------------- 9.4/14.8 MB 2.0 MB/s eta 0:00:03\n", " --------------------------- ------------ 10.2/14.8 MB 2.1 MB/s eta 0:00:03\n", " ----------------------------- ---------- 10.7/14.8 MB 2.1 MB/s eta 0:00:02\n", " ------------------------------ --------- 11.3/14.8 MB 2.1 MB/s eta 0:00:02\n", " ------------------------------- -------- 11.5/14.8 MB 2.1 MB/s eta 0:00:02\n", " -------------------------------- ------- 12.1/14.8 MB 2.1 MB/s eta 0:00:02\n", " --------------------------------- ------ 12.3/14.8 MB 2.1 MB/s eta 0:00:02\n", " --------------------------------- ------ 12.6/14.8 MB 2.1 MB/s eta 0:00:02\n", " ----------------------------------- ---- 13.1/14.8 MB 2.1 MB/s eta 0:00:01\n", " ------------------------------------ --- 13.4/14.8 MB 2.1 MB/s eta 0:00:01\n", " ------------------------------------- -- 13.9/14.8 MB 2.0 MB/s eta 0:00:01\n", " -------------------------------------- - 14.2/14.8 MB 2.0 MB/s eta 0:00:01\n", " --------------------------------------- 14.7/14.8 MB 2.0 MB/s eta 0:00:01\n", " ---------------------------------------- 14.8/14.8 MB 2.0 MB/s eta 0:00:00\n", "Downloading narwhals-1.38.2-py3-none-any.whl (338 kB)\n", "Installing collected packages: narwhals, plotly\n", "Successfully installed narwhals-1.38.2 plotly-6.0.1\n" ] } ], "source": [ "#!pip install plotly\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "qVkh_Y7HWj9w", "metadata": { "executionInfo": { "elapsed": 10133, "status": "ok", "timestamp": 1746448329851, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "qVkh_Y7HWj9w" }, "outputs": [], "source": [ "import pandas as pd\n", "import jieba\n", "import jieba.analyse\n", "import re\n", "import numpy as np\n", "from collections import defaultdict\n", "import multiprocessing\n", "\n", "from gensim.models.phrases import Phrases, Phraser\n", "from gensim.models import Word2Vec, KeyedVectors\n", "\n", "from gensim.models import Word2Vec\n", "\n", "import matplotlib.pyplot as plt\n", "\n", "from sklearn.decomposition import PCA\n", "from sklearn.manifold import TSNE\n", "\n", "import seaborn as sns\n", "import torch\n", "\n", "from matplotlib.font_manager import fontManager\n", "import plotly.express as px\n", "\n", "sns.set_style(\"darkgrid\")" ] }, { "cell_type": "code", "execution_count": null, "id": "076c6263", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 5331, "status": "ok", "timestamp": 1746448337568, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "076c6263", "outputId": "20be4efe-6dc5-4f3c-be04-3527cfe7323c" }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Building prefix dict from c:\\Users\\rolya\\Desktop\\DIV\\divoce_project2\\dict\\dict.txt.big ...\n", "Dumping model to file cache C:\\Users\\rolya\\AppData\\Local\\Temp\\jieba.u2ecd0c6dc6535871c5cc6cd50f65ab67.cache\n", "Loading model cost 1.600 seconds.\n", "Prefix dict has been built successfully.\n" ] } ], "source": [ "# 設定繁體中文詞庫\n", "jieba.set_dictionary('/Users/rolya/Desktop/DIV/divoce_project2/dict/dict.txt.big')\n", "jieba.load_userdict('/Users/rolya/Desktop/DIV/divoce_project2/dict/user_dict.txt')\n", "\n", "# 新增stopwords\n", "with open('/Users/rolya/Desktop/DIV/divoce_project2/dict/stopwords.txt',encoding=\"utf-8\") as f:\n", " stopWords = [line.strip() for line in f.readlines()]" ] }, { "cell_type": "code", "execution_count": 4, "id": "CLy-xuUYW8Mb", "metadata": { "executionInfo": { "elapsed": 19, "status": "ok", "timestamp": 1746448340493, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "CLy-xuUYW8Mb" }, "outputs": [], "source": [ "# 斷詞函式\n", "def getToken(row):\n", " seg_list = jieba.lcut(row)\n", " seg_list = [w for w in seg_list if w not in stopWords and len(w)>1] # 篩選掉停用字與字元數小於1的詞彙\n", "\n", " return seg_list" ] }, { "cell_type": "code", "execution_count": 5, "id": "UH4Gyx2zW8O3", "metadata": { "executionInfo": { "elapsed": 5350, "status": "ok", "timestamp": 1746448346821, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "UH4Gyx2zW8O3" }, "outputs": [], "source": [ "# 讀入中文示範資料集\n", "origin_data = pd.read_csv('/Users/rolya/Desktop/DIV/divoce_project2/text.csv')" ] }, { "cell_type": "code", "execution_count": 6, "id": "I1D6HJoyW8RK", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 1000 }, "executionInfo": { "elapsed": 55259, "status": "ok", "timestamp": 1746448404395, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "I1D6HJoyW8RK", "outputId": "faaa04de-fe22-4611-d081-de1d5f4fefb5" }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartUrlartTitleartDateartContentsentenceword
11https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...偷看手機是不對的[偷看, 手機]
21https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...但如果已經結婚了[結婚]
31https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...太太想看你手機[太太, 手機]
41https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...真的可以拒絕嗎[真的, 拒絕]
51https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...感覺你拒絕[感覺, 拒絕]
61https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...就是心裡有鬼[有鬼]
71https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...只是讓太太猜忌[太太, 猜忌]
81https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...自己日子難過[日子, 難過]
101https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...我手機都是隨便太太看[手機, 隨便, 太太]
111https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...但每次都被看的提心吊膽[每次, 提心吊膽]
\n", "
" ], "text/plain": [ " system_id artUrl \\\n", "1 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "2 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "3 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "4 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "5 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "6 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "7 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "8 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "10 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "11 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "\n", " artTitle artDate \\\n", "1 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "2 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "3 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "4 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "5 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "6 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "7 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "8 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "10 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "11 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "\n", " artContent sentence \\\n", "1 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 偷看手機是不對的 \n", "2 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 但如果已經結婚了 \n", "3 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 太太想看你手機 \n", "4 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 真的可以拒絕嗎 \n", "5 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 感覺你拒絕 \n", "6 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 就是心裡有鬼 \n", "7 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 只是讓太太猜忌 \n", "8 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 自己日子難過 \n", "10 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 我手機都是隨便太太看 \n", "11 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 但每次都被看的提心吊膽 \n", "\n", " word \n", "1 [偷看, 手機] \n", "2 [結婚] \n", "3 [太太, 手機] \n", "4 [真的, 拒絕] \n", "5 [感覺, 拒絕] \n", "6 [有鬼] \n", "7 [太太, 猜忌] \n", "8 [日子, 難過] \n", "10 [手機, 隨便, 太太] \n", "11 [每次, 提心吊膽] " ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 資料前處理\n", "\n", "# 去除一些不需要的欄位\n", "metaData = origin_data.drop(['artPoster', 'artCatagory', 'artComment', 'e_ip', 'insertedDate', 'dataSource'], axis=1)\n", "\n", "# 只留下中文字\n", "metaData['sentence'] = metaData['artContent'].str.replace(r'\\n\\n','。', regex=True)\n", "metaData['sentence'] = metaData['sentence'].str.replace(r'\\n','', regex=True)\n", "\n", "metaData['sentence'] = metaData['sentence'].str.split(\"[,，。！!？?]{1,}\")\n", "metaData = metaData.explode('sentence').reset_index(drop=True)\n", "\n", "metaData['sentence'] = metaData['sentence'].apply(lambda x: re.sub('[^\\u4e00-\\u9fff]+', '',x))\n", "\n", "metaData['word'] = metaData.sentence.apply(getToken)\n", "\n", "metaData = metaData[metaData['word'].apply(len) > 0]\n", "\n", "metaData.head(10)" ] }, { "cell_type": "code", "execution_count": 7, "id": "8jPUK7SdXSqx", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 920 }, "executionInfo": { "elapsed": 1755, "status": "ok", "timestamp": 1746448413451, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "8jPUK7SdXSqx", "outputId": "10e01979-80cb-49d2-f737-1eb34a779264" }, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartUrlartTitleartDateartContentsentencewordword_list_bigrams
11https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...偷看手機是不對的[偷看, 手機][偷看, 手機]
21https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...但如果已經結婚了[結婚][結婚]
31https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...太太想看你手機[太太, 手機][太太, 手機]
41https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...真的可以拒絕嗎[真的, 拒絕][真的, 拒絕]
51https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...感覺你拒絕[感覺, 拒絕][感覺, 拒絕]
\n", "
" ], "text/plain": [ " system_id artUrl \\\n", "1 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "2 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "3 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "4 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "5 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "\n", " artTitle artDate \\\n", "1 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "2 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "3 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "4 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "5 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "\n", " artContent sentence word \\\n", "1 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 偷看手機是不對的 [偷看, 手機] \n", "2 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 但如果已經結婚了 [結婚] \n", "3 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 太太想看你手機 [太太, 手機] \n", "4 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 真的可以拒絕嗎 [真的, 拒絕] \n", "5 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 感覺你拒絕 [感覺, 拒絕] \n", "\n", " word_list_bigrams \n", "1 [偷看, 手機] \n", "2 [結婚] \n", "3 [太太, 手機] \n", "4 [真的, 拒絕] \n", "5 [感覺, 拒絕] " ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "sents = metaData['word'].to_list()\n", "bigrams = Phrases(sents,min_count=1, threshold=1000)\n", "bigram_phrasers = Phraser(bigrams)\n", "metaData['word_list_bigrams'] = list(bigram_phrasers[sents])\n", "\n", "metaData.head()" ] }, { "cell_type": "code", "execution_count": 8, "id": "dzOfLkazXbES", "metadata": { "executionInfo": { "elapsed": 151, "status": "ok", "timestamp": 1746448416220, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "dzOfLkazXbES" }, "outputs": [], "source": [ "word_freq = defaultdict(int)\n", "# 計算詞頻\n", "sents = metaData['word_list_bigrams'].tolist()\n", "for sent in sents: # sent 中的每個句子\n", " for i in sent: # i 是句子中的每個字\n", " word_freq[i] += 1" ] }, { "cell_type": "code", "execution_count": 9, "id": "BadyXcj5XbG6", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 12, "status": "ok", "timestamp": 1746448417825, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "BadyXcj5XbG6", "outputId": "e53931fe-27d1-45c3-9c05-f8b2c01b112f" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "total unique words in sentences: 39393\n" ] }, { "data": { "text/plain": [ "['小孩', '離婚', '老婆', '老公', '真的', '孩子', '工作', '婚姻', '結婚', '太太']" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "print(f\"total unique words in sentences: {len(word_freq)}\")\n", "sorted(word_freq, key=word_freq.get, reverse=True)[:10]" ] }, { "cell_type": "code", "execution_count": 10, "id": "bQxIwrRPXp8L", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 27, "status": "ok", "timestamp": 1746448439640, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "bQxIwrRPXp8L", "outputId": "43c3a5e7-bf56-47e5-ccd8-5bcc39805e1e" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "sentence number of corpus: 95013\n", "average length of sentences: 3.452327576226411\n" ] } ], "source": [ "print(f\"sentence number of corpus: {len(sents)}\")\n", "i = 0\n", "for sent in sents:\n", " i = i + len(sent)\n", "print(f\"average length of sentences: {i/len(sents)}\")" ] }, { "cell_type": "code", "execution_count": 11, "id": "M8JXl54lX1B8", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 14, "status": "ok", "timestamp": 1746448441412, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "M8JXl54lX1B8", "outputId": "4ffee132-4a9b-44f2-c8cc-0bc45217870d" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "env: PYTHONHASHSEED=2025\n" ] } ], "source": [ "# 環境變數設定\n", "%env PYTHONHASHSEED=2025" ] }, { "cell_type": "code", "execution_count": 12, "id": "-dlYF-HbX1Ew", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 10, "status": "ok", "timestamp": 1746448442578, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "-dlYF-HbX1Ew", "outputId": "4d6246f8-a6e3-494e-af2d-e0f9d8386f7f" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "number of cores: 8\n" ] } ], "source": [ "# 查看機器的core\n", "cores = multiprocessing.cpu_count()\n", "print(f\"number of cores: {cores}\")" ] }, { "cell_type": "code", "execution_count": 13, "id": "HYUphL5uX1Gx", "metadata": { "executionInfo": { "elapsed": 14626, "status": "ok", "timestamp": 1746448458777, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "HYUphL5uX1Gx" }, "outputs": [], "source": [ "# 建立模型\n", "w2v_model = Word2Vec(sents,\n", " min_count=30,# 小於30次tf的字會被刪除\n", " window=2,# 往左右各2的距離\n", " vector_size=128,# vector 的維度\n", " sample=0.005,# 愈小的話，高tf的字會不容易被選到\n", " alpha=0.001,# learning rate\n", " min_alpha=0.0005, # 迭代到最小的learning rate，learning rate會慢慢下降至min_alpha\n", " negative=0,\n", " workers=cores-1, # 用的cpu資源\n", " seed=8787,\n", " sg = 1,# 0/1 是否使用skip gram\n", " epochs= 30,\n", " hs=1 , # hierarchical softmax\n", " )" ] }, { "cell_type": "code", "execution_count": 14, "id": "F4aco4pYX1I6", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 9, "status": "ok", "timestamp": 1746448460904, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "F4aco4pYX1I6", "outputId": "db21bd03-797e-4505-a82b-fed12df4b761" }, "outputs": [ { "data": { "text/plain": [ "[('思考', 0.7755464315414429),\n", " ('解決', 0.7708988785743713),\n", " ('期待', 0.6676611304283142),\n", " ('嘗試', 0.6662196516990662),\n", " ('這是', 0.6276875734329224),\n", " ('衝突', 0.6261622309684753),\n", " ('表達', 0.6218885183334351),\n", " ('事件', 0.6023396253585815),\n", " ('想法', 0.5949125289916992),\n", " ('需求', 0.5907225012779236)]" ] }, "execution_count": 14, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 檢查最相關的字\n", "w2v_model.wv.most_similar('溝通',topn=10)" ] }, { "cell_type": "code", "execution_count": 15, "id": "DQBpLVepX-eW", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 10, "status": "ok", "timestamp": 1746448462758, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "DQBpLVepX-eW", "outputId": "7f0f0ac5-b93e-422f-bd9b-02aeb36c0933" }, "outputs": [ { "data": { "text/plain": [ "[('提出', 0.8008367419242859),\n", " ('原因', 0.7528184652328491),\n", " ('不爽', 0.7414447665214539),\n", " ('分手', 0.7389088869094849),\n", " ('實在', 0.732452392578125),\n", " ('念頭', 0.7268778681755066),\n", " ('出軌', 0.7090317606925964),\n", " ('乾脆', 0.700692892074585),\n", " ('平靜', 0.693835437297821),\n", " ('理由', 0.6839743852615356)]" ] }, "execution_count": 15, "metadata": {}, "output_type": "execute_result" } ], "source": [ "w2v_model.wv.most_similar('外遇',topn=10)" ] }, { "cell_type": "code", "execution_count": 16, "id": "TGj28f87X-gx", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 44, "status": "ok", "timestamp": 1746448464700, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "TGj28f87X-gx", "outputId": "551f9ab0-5c20-48ea-e59b-aab175a35b6c" }, "outputs": [ { "data": { "text/plain": [ "[('平靜', 0.7731956839561462),\n", " ('解決', 0.7368565797805786),\n", " ('思考', 0.7182171940803528),\n", " ('值得', 0.715507447719574),\n", " ('理由', 0.70955491065979),\n", " ('想法', 0.6994882225990295),\n", " ('提出', 0.6934615969657898),\n", " ('尊重', 0.6934552788734436),\n", " ('道理', 0.6815539598464966),\n", " ('原因', 0.6742791533470154)]" ] }, "execution_count": 16, "metadata": {}, "output_type": "execute_result" } ], "source": [ "w2v_model.wv.most_similar(['溝通','外遇'],topn=10)" ] }, { "cell_type": "code", "execution_count": 17, "id": "Ah-Z2aBOYTWu", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 8, "status": "ok", "timestamp": 1746448466738, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "Ah-Z2aBOYTWu", "outputId": "2224e8fc-3314-4fee-e707-8363e1c0be2f" }, "outputs": [ { "data": { "text/plain": [ "[('休息', 0.6033270359039307),\n", " ('白天', 0.5987855792045593),\n", " ('在家', 0.5936435461044312),\n", " ('大人', 0.5754421353340149),\n", " ('足夠', 0.5643977522850037),\n", " ('接送', 0.5636379718780518),\n", " ('下班', 0.5620827078819275),\n", " ('心力', 0.5549818277359009),\n", " ('育兒', 0.5517842173576355),\n", " ('旁邊', 0.5506572723388672)]" ] }, "execution_count": 17, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 跟兩個字最不相關\n", "w2v_model.wv.most_similar(negative=['外遇','離婚'],topn=10)" ] }, { "cell_type": "code", "execution_count": 18, "id": "ZYb3cV1GYdn-", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 8, "status": "ok", "timestamp": 1746448472457, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "ZYb3cV1GYdn-", "outputId": "97a5fe78-5dd4-4e1a-d718-76489e0a53bb" }, "outputs": [ { "data": { "text/plain": [ "-0.18695961" ] }, "execution_count": 18, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 計算兩個字之間的關係\n", "w2v_model.wv.similarity(\"財產\",\"家庭\")" ] }, { "cell_type": "code", "execution_count": 19, "id": "rfT2fDmZYTZZ", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 8, "status": "ok", "timestamp": 1746448509442, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "rfT2fDmZYTZZ", "outputId": "81522c70-ce3a-4ab1-f006-34c71752ff24" }, "outputs": [ { "data": { "text/plain": [ "0.55098337" ] }, "execution_count": 19, "metadata": {}, "output_type": "execute_result" } ], "source": [ "w2v_model.wv.similarity(\"孩子\",\"照顧\")" ] }, { "cell_type": "code", "execution_count": 20, "id": "ap-85hspYw2O", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 35 }, "executionInfo": { "elapsed": 22, "status": "ok", "timestamp": 1746448563929, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "ap-85hspYw2O", "outputId": "a3a5955f-62ea-4241-ff21-f83790a42027" }, "outputs": [ { "data": { "text/plain": [ "'財產'" ] }, "execution_count": 20, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 比較字詞間，誰最不相關（邊緣）\n", "w2v_model.wv.doesnt_match([\"孩子\", \"照顧\", '財產'])" ] }, { "cell_type": "code", "execution_count": 21, "id": "4nJ3n4dUYw4s", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 9, "status": "ok", "timestamp": 1746448571806, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "4nJ3n4dUYw4s", "outputId": "fe67bec2-9262-4630-bfd6-dc8499cc8790" }, "outputs": [ { "data": { "text/plain": [ "[('痛苦', 0.5897860527038574),\n", " ('故事', 0.5843417048454285),\n", " ('紀錄', 0.5227385759353638),\n", " ('妻子', 0.5202898979187012),\n", " ('個性', 0.5193036198616028)]" ] }, "execution_count": 21, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 相對關係\n", "w2v_model.wv.most_similar(positive=[\"孩子\"], negative=[\"照顧\"], topn=5)" ] }, { "cell_type": "code", "execution_count": 22, "id": "bh-G_76TZFo8", "metadata": { "executionInfo": { "elapsed": 3, "status": "ok", "timestamp": 1746448575071, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "bh-G_76TZFo8" }, "outputs": [], "source": [ "# 取得所有的字\n", "words = w2v_model.wv.key_to_index.keys()" ] }, { "cell_type": "code", "execution_count": 23, "id": "vMtqPxNDZFrQ", "metadata": { "executionInfo": { "elapsed": 3, "status": "ok", "timestamp": 1746448577427, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "vMtqPxNDZFrQ" }, "outputs": [], "source": [ "# 降維：利用PCA tSNE\n", "\n", "def reduceDim(mat,method:str='PCA',dim:str=2,perplexity = 25,learning_rate = 400):\n", "\n", " method_dict = {\n", " \"PCA\":PCA(n_components=dim,iterated_power = 1000,random_state=0),\n", " \"TSNE\":TSNE(n_components=dim,random_state=0,perplexity=perplexity,learning_rate=learning_rate),\n", " }\n", " new_feat = method_dict[method].fit_transform(mat)\n", "\n", " return new_feat" ] }, { "cell_type": "code", "execution_count": 24, "id": "RKMymLiIZL_6", "metadata": { "executionInfo": { "elapsed": 4, "status": "ok", "timestamp": 1746448579681, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "RKMymLiIZL_6" }, "outputs": [], "source": [ "# 拿到list of words 的vector\n", "def getVecs(model,words:list):\n", " vecs = []\n", " for i in words:\n", " vecs.append(model.wv[i])\n", " return np.vstack(vecs)" ] }, { "cell_type": "code", "execution_count": 25, "id": "IhkKOa7nZOoC", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 10, "status": "ok", "timestamp": 1746448582558, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "IhkKOa7nZOoC", "outputId": "ee073901-e574-432c-f907-3b177cda3c32" }, "outputs": [ { "data": { "text/plain": [ "array([[ 0.05552926, 0.03847976, -0.23192777, 0.06041394, -0.03632559,\n", " -0.04825368, -0.09273008, 0.06537188, 0.07540878, 0.11345442,\n", " 0.10255916, -0.077728 , 0.03348193, -0.02929123, 0.04965089,\n", " 0.04994745, -0.02007875, 0.06827826, -0.00043407, -0.01652355,\n", " 0.07447162, 0.06956953, 0.04445267, -0.06862568, -0.01181999,\n", " -0.03680974, -0.09098391, -0.02162968, 0.0925396 , -0.03726147,\n", " 0.03511068, -0.0015918 , -0.04098483, -0.08775292, 0.02748766,\n", " -0.0189511 , 0.09938143, -0.00548935, -0.16770932, 0.03284688,\n", " 0.05587117, -0.09926225, -0.09662215, 0.02078925, -0.0978763 ,\n", " -0.10456892, -0.08619943, -0.07782117, -0.00280136, -0.02179379,\n", " 0.10193834, 0.0503955 , -0.00199134, 0.06327718, 0.05917658,\n", " -0.00107532, -0.03456983, -0.12342957, 0.02685144, -0.00024162,\n", " -0.08171521, 0.02174407, 0.03526353, -0.0224024 , 0.01412679,\n", " -0.03493559, -0.07929365, -0.11893469, -0.07531871, 0.04449066,\n", " -0.10216135, 0.10212398, -0.09488969, 0.03615135, -0.03160409,\n", " 0.10345982, 0.03115953, 0.06215408, -0.1881784 , -0.03869899,\n", " 0.0145611 , 0.08423696, -0.00474182, 0.01020931, -0.05811653,\n", " -0.08193173, -0.09165919, 0.00748084, -0.01750713, 0.02589018,\n", " -0.07287675, 0.05250846, 0.02563178, -0.08972724, -0.02054487,\n", " -0.03332768, 0.07814557, -0.10100207, 0.00216089, 0.13747457,\n", " 0.07158327, 0.02981087, 0.02596068, 0.00235181, 0.01135799,\n", " 0.03610549, -0.03554225, 0.1099498 , -0.07154053, 0.02411028,\n", " 0.10968643, -0.0749239 , 0.1138766 , 0.01662243, 0.00771771,\n", " -0.07943476, -0.06997006, -0.04681144, -0.02465238, -0.04719803,\n", " 0.09851346, -0.05417875, 0.05610305, -0.16223566, -0.08995478,\n", " 0.08608418, 0.057287 , -0.014308 ],\n", " [ 0.06940024, 0.02854187, 0.0235089 , 0.0113638 , -0.05424808,\n", " -0.03968925, 0.01986084, 0.04322903, -0.04480281, -0.02419144,\n", " 0.08776139, -0.07677491, 0.03418854, 0.04797035, 0.02194501,\n", " 0.08888641, -0.02592358, 0.00501392, 0.0174946 , -0.02264285,\n", " -0.00230185, -0.0468002 , 0.01820106, 0.04159217, -0.02537901,\n", " -0.05261362, 0.00505259, 0.00644564, -0.01651222, -0.04771264,\n", " -0.03318613, 0.01206908, -0.00102497, 0.05461093, 0.06607899,\n", " -0.005135 , 0.06350551, 0.07255646, -0.01323139, 0.00854335,\n", " -0.02009638, 0.03134184, -0.03179 , -0.04898661, -0.07075465,\n", " -0.05226701, 0.00413447, 0.05127696, -0.04384896, -0.05382123,\n", " 0.05206602, 0.04317016, -0.02729745, 0.03092107, 0.00511999,\n", " 0.02858668, -0.0191604 , -0.05599732, -0.02721906, 0.01753448,\n", " 0.04740141, -0.00360852, 0.01634916, -0.07158028, -0.04917749,\n", " -0.01737433, -0.01424111, -0.02998464, 0.03152172, 0.08365235,\n", " -0.02340383, 0.05740769, 0.01968239, -0.02477934, 0.03875388,\n", " 0.05084869, 0.00639911, 0.03731444, -0.06243524, 0.00063726,\n", " -0.07853391, 0.02793852, -0.03870003, 0.00755745, 0.0322638 ,\n", " -0.02977583, -0.02146657, 0.05491862, 0.02703132, -0.04046306,\n", " 0.04141882, -0.019448 , -0.03544722, 0.05065524, 0.02812297,\n", " -0.00278463, 0.00866573, -0.02489917, 0.0713686 , 0.01389261,\n", " 0.00163449, 0.05256031, -0.03146213, -0.02890553, -0.00148923,\n", " -0.02173272, 0.06234517, 0.07063311, 0.01675576, 0.04644315,\n", " -0.01290863, -0.04317336, 0.06099079, -0.04302061, -0.02277633,\n", " -0.02732737, 0.03801377, -0.0165253 , 0.00495038, -0.04724422,\n", " -0.00122255, 0.05922864, -0.0054045 , -0.01885483, 0.03823422,\n", " 0.01087835, -0.01277172, 0.05238632]], dtype=float32)" ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "getVecs(w2v_model,['溝通','外遇'])" ] }, { "cell_type": "code", "execution_count": 26, "id": "4DxkIg3rZTpX", "metadata": { "executionInfo": { "elapsed": 8, "status": "ok", "timestamp": 1746448593237, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "4DxkIg3rZTpX" }, "outputs": [], "source": [ "# 擴展相似的字詞\n", "def expandPosWord(model, words:list, top_n:int, split = True):\n", "\n", " if split == False:\n", " wp = model.wv.most_similar(words,topn = top_n)\n", " return wp\n", " expand = []\n", "\n", " for w in words:\n", " wp = model.wv.most_similar(w,topn = top_n)\n", " for i in wp:\n", " expand.append(i[0])\n", "\n", " return list(set(expand))" ] }, { "cell_type": "code", "execution_count": 27, "id": "0VpyujvPZTr7", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 7, "status": "ok", "timestamp": 1746448594941, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "0VpyujvPZTr7", "outputId": "b58643d0-c102-4e48-b07f-fc55aa8f4cb9" }, "outputs": [ { "data": { "text/plain": [ "['期待',\n", " '解決',\n", " '提出',\n", " '實在',\n", " '表達',\n", " '想法',\n", " '乾脆',\n", " '需求',\n", " '事件',\n", " '原因',\n", " '嘗試',\n", " '思考',\n", " '分手',\n", " '理由',\n", " '出軌',\n", " '衝突',\n", " '不爽',\n", " '這是',\n", " '平靜',\n", " '念頭']" ] }, "execution_count": 27, "metadata": {}, "output_type": "execute_result" } ], "source": [ "expandPosWord(w2v_model,['溝通','外遇'],top_n = 10)" ] }, { "cell_type": "markdown", "id": "gLuu-WhtZmg-", "metadata": { "id": "gLuu-WhtZmg-" }, "source": [ "以上字詞是依照第二次專案tf-idf前十大字詞\n", "4828\t工作\n", "3219\t問題\n", "9604\t美國\n", "4481\t家庭\n", "10732\t財產\n", "7951\t溝通\n", "3646\t外遇\n", "6672\t改變\n", "8172\t照顧\n", "1882\t公公" ] }, { "cell_type": "code", "execution_count": 34, "id": "K6ldPzJtaIWl", "metadata": { "executionInfo": { "elapsed": 3, "status": "ok", "timestamp": 1746448627096, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "K6ldPzJtaIWl" }, "outputs": [], "source": [ "# 畫出兩維的散佈圖\n", "def plotScatter(vec_df):\n", " \"\"\"\n", " vec_df: 字詞及其兩個維度的值\n", " \"\"\"\n", " plt.figure(figsize=(15,15))\n", " fontManager.addfont('/Users/rolya/Desktop/DIV/divoce_project2/TaipeiSansTCBeta-Regular.ttf')\n", " plt.rcParams['font.sans-serif'] = ['Taipei Sans TC Beta']\n", " plt.rcParams['font.size'] = '16'\n", "\n", " p = sns.scatterplot(x=\"dim1\", y=\"dim2\",\n", " data=vec_df)\n", " for line in range(0, vec_df.shape[0]):\n", " p.text(vec_df[\"dim1\"][line],\n", " vec_df['dim2'][line],\n", " ' ' + vec_df[\"word\"][line].title(),\n", " horizontalalignment='left',\n", " verticalalignment='bottom', size='medium',\n", " weight='normal'\n", " ).set_size(15)\n", " plt.show()\n", "\n", "# 畫出三維的散佈圖\n", "def plotScatter3D(vec_df):\n", " vec_df['size'] = .5\n", " if 'color' not in vec_df.columns:\n", " vec_df['color'] = 'blue'\n", " fig = px.scatter_3d(\n", " vec_df,'dim1','dim2','dim3',text = 'word',width=800, height=800,color = 'color',size = 'size'\n", "\n", " )\n", "\n", " fig.show()" ] }, { "cell_type": "code", "execution_count": 35, "id": "ZnaUyfFKaIZa", "metadata": { "colab": { "base_uri": "https://localhost:8080/" }, "executionInfo": { "elapsed": 1058, "status": "ok", "timestamp": 1746449496210, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "ZnaUyfFKaIZa", "outputId": "bb1d6728-6674-48fc-82c6-bb86a4bec9bb" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(150, 128)\n", "(150, 2)\n" ] } ], "source": [ "sample_words = np.random.choice(list(words),150,replace=False).tolist()\n", "\n", "feat = getVecs(model=w2v_model,words=sample_words)\n", "print(feat.shape)\n", "new_feat = reduceDim(feat,method='TSNE',perplexity=20,learning_rate = 800)\n", "print(new_feat.shape)" ] }, { "cell_type": "code", "execution_count": 36, "id": "kP2XUbI_bIUj", "metadata": { "executionInfo": { "elapsed": 11, "status": "ok", "timestamp": 1746449496791, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "kP2XUbI_bIUj" }, "outputs": [], "source": [ "word_df = pd.DataFrame({\n", " \"word\":sample_words,\n", " \"dim1\":new_feat[:,0],\n", " \"dim2\":new_feat[:,1],\n", "})" ] }, { "cell_type": "code", "execution_count": 37, "id": "-falARSubKP7", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 940 }, "executionInfo": { "elapsed": 559, "status": "ok", "timestamp": 1746449499304, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "-falARSubKP7", "outputId": "e80f6271-5376-47ab-b5f1-d5bcd331f1a4" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plotScatter(word_df)" ] }, { "cell_type": "code", "execution_count": 39, "id": "mBuzpve_bMFr", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 834 }, "executionInfo": { "elapsed": 88, "status": "ok", "timestamp": 1746449505942, "user": { "displayName": "章茗鈞", "userId": "09803695490438841130" }, "user_tz": -480 }, "id": "mBuzpve_bMFr", "outputId": "088b6722-7c64-430a-8ab2-a99da4bac2bd" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "(150, 3)\n" ] }, { "data": { "application/vnd.plotly.v1+json": { "config": { "plotlyServerURL": "https://plot.ly" }, "data": [ { "hovertemplate": "color=blue
dim1=%{x}
dim2=%{y}
dim3=%{z}
size=%{marker.size}
word=%{text}", "legendgroup": "blue", "marker": { "color": "#636efa", "size": { "bdata": "AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/", "dtype": "f8" }, "sizemode": "area", "sizeref": 0.00125, "symbol": "circle" }, "mode": "markers+text", "name": "blue", "scene": "scene", "showlegend": true, "text": [ "抱抱", "扭曲", "禮拜", "實際上", "扣掉", "例子", "男友", "證明", "好好", "消失", "可能性", "打算", "碰到", "分配", "提早", "諮詢", "相關", "照片", "專業", "我能", "更是", "老公", "依舊", "不在乎", "吞下去", "勇敢", "體力", "那位", "墮胎", "假日", "本質", "對話", "法官", "電動", "單方面", "支撐", "這不", "同事", "努力", "講好", "在乎", "項目", "婚姻生活", "居住", "婆媳", "工廠", "生完", "分手", "互動", "要生", "律師", "敘述", "賣掉", "沒差", "停止", "安全感", "盡力", "繼承", "透天", "盡量", "生產", "三次", "陌生人", "老二", "沒事", "來回", "心疼", "發現", "安慰", "所有人", "留言", "表面", "道歉", "接近", "扶養", "頭期", "謝謝", "上次", "第一個", "偷吃", "貸款", "大小事", "男人", "有用", "身邊", "男生", "反駁", "阿公", "無限", "背叛", "困難", "分析", "做出", "丈夫", "未婚夫", "奶瓶", "脾氣", "義務", "你好", "無所謂", "花費", "心情", "做好", "情境", "這份", "拿到", "電視", "養育", "北京", "參與", "扶持", "人才", "犧牲", "失望", "財產", "一路", "觀念", "最終", "危險", "虐待", "怨恨", "改善", "釐清", "磨合", "言論", "情緒", "部份", "我並", "女同事", "提供", "感謝", "贍養費", "打開", "一步", "婚姻", "做法", "姐姐", "言語", "分擔", "長時間", "兩人", "作法", "當你", "拿出", "走路", "撫養費", "伴侶", "管理", "睡眠", "住家" ], "type": "scatter3d", "x": { "bdata": "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", "dtype": "f4" }, "y": { "bdata": "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", "dtype": "f4" }, "z": { "bdata": "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", "dtype": "f4" } } ], "layout": { "height": 800, "legend": { "itemsizing": "constant", "title": { "text": "color" }, "tracegroupgap": 0 }, "margin": { "t": 60 }, "scene": { "domain": { "x": [ 0, 1 ], "y": [ 0, 1 ] }, "xaxis": { "title": { "text": "dim1" } }, "yaxis": { "title": { "text": "dim2" } }, "zaxis": { "title": { "text": "dim3" } } }, "template": { "data": { "bar": [ { "error_x": { "color": "#2a3f5f" }, "error_y": { "color": "#2a3f5f" }, "marker": { "line": { "color": "#E5ECF6", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "bar" } ], "barpolar": [ { "marker": { "line": { "color": "#E5ECF6", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "barpolar" } ], "carpet": [ { "aaxis": { "endlinecolor": "#2a3f5f", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "#2a3f5f" }, "baxis": { "endlinecolor": "#2a3f5f", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "#2a3f5f" }, "type": "carpet" } ], "choropleth": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "choropleth" } ], "contour": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "contour" } ], "contourcarpet": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "contourcarpet" } ], "heatmap": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "heatmap" } ], "histogram": [ { "marker": { "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "histogram" } ], "histogram2d": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "histogram2d" } ], "histogram2dcontour": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "histogram2dcontour" } ], "mesh3d": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "mesh3d" } ], "parcoords": [ { "line": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "parcoords" } ], "pie": [ { "automargin": true, "type": "pie" } ], "scatter": [ { "fillpattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 }, "type": "scatter" } ], "scatter3d": [ { "line": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatter3d" } ], "scattercarpet": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattercarpet" } ], "scattergeo": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattergeo" } ], "scattergl": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattergl" } ], "scattermap": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattermap" } ], "scattermapbox": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattermapbox" } ], "scatterpolar": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterpolar" } ], "scatterpolargl": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterpolargl" } ], "scatterternary": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterternary" } ], "surface": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "surface" } ], "table": [ { "cells": { "fill": { "color": "#EBF0F8" }, "line": { "color": "white" } }, "header": { "fill": { "color": "#C8D4E3" }, "line": { "color": "white" } }, "type": "table" } ] }, "layout": { "annotationdefaults": { "arrowcolor": "#2a3f5f", "arrowhead": 0, "arrowwidth": 1 }, "autotypenumbers": "strict", "coloraxis": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "colorscale": { "diverging": [ [ 0, "#8e0152" ], [ 0.1, "#c51b7d" ], [ 0.2, "#de77ae" ], [ 0.3, "#f1b6da" ], [ 0.4, "#fde0ef" ], [ 0.5, "#f7f7f7" ], [ 0.6, "#e6f5d0" ], [ 0.7, "#b8e186" ], [ 0.8, "#7fbc41" ], [ 0.9, "#4d9221" ], [ 1, "#276419" ] ], "sequential": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "sequentialminus": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ] }, "colorway": [ "#636efa", "#EF553B", "#00cc96", "#ab63fa", "#FFA15A", "#19d3f3", "#FF6692", "#B6E880", "#FF97FF", "#FECB52" ], "font": { "color": "#2a3f5f" }, "geo": { "bgcolor": "white", "lakecolor": "white", "landcolor": "#E5ECF6", "showlakes": true, "showland": true, "subunitcolor": "white" }, "hoverlabel": { "align": "left" }, "hovermode": "closest", "mapbox": { "style": "light" }, "paper_bgcolor": "white", "plot_bgcolor": "#E5ECF6", "polar": { "angularaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "bgcolor": "#E5ECF6", "radialaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" } }, "scene": { "xaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" }, "yaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" }, "zaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" } }, "shapedefaults": { "line": { "color": "#2a3f5f" } }, "ternary": { "aaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "baxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "bgcolor": "#E5ECF6", "caxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" } }, "title": { "x": 0.05 }, "xaxis": { "automargin": true, "gridcolor": "white", "linecolor": "white", "ticks": "", "title": { "standoff": 15 }, "zerolinecolor": "white", "zerolinewidth": 2 }, "yaxis": { "automargin": true, "gridcolor": "white", "linecolor": "white", "ticks": "", "title": { "standoff": 15 }, "zerolinecolor": "white", "zerolinewidth": 2 } } }, "width": 800 } } }, "metadata": {}, "output_type": "display_data" } ], "source": [ "new_feat = reduceDim(feat,dim = 3,method = 'PCA' )\n", "print(new_feat.shape)\n", "word_df = pd.DataFrame({\n", " \"word\":sample_words,\n", " \"dim1\":new_feat[:,0],\n", " \"dim2\":new_feat[:,1],\n", " \"dim3\":new_feat[:,2],\n", "})\n", "plotScatter3D(word_df)" ] }, { "cell_type": "markdown", "id": "UjwOlPoXdjjh", "metadata": { "id": "UjwOlPoXdjjh" }, "source": [ "3D圖能縮放" ] }, { "cell_type": "markdown", "id": "bc846a67", "metadata": {}, "source": [ "將字分群\n" ] }, { "cell_type": "code", "execution_count": null, "id": "c5fe440a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Collecting scikit-learn-extra\n", " Downloading scikit_learn_extra-0.3.0-cp311-cp311-win_amd64.whl.metadata (3.7 kB)\n", "Requirement already satisfied: numpy>=1.13.3 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn-extra) (1.24.3)\n", "Requirement already satisfied: scipy>=0.19.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn-extra) (1.12.0)\n", "Requirement already satisfied: scikit-learn>=0.23.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn-extra) (1.4.0)\n", "Requirement already satisfied: joblib>=1.2.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn>=0.23.0->scikit-learn-extra) (1.4.2)\n", "Requirement already satisfied: threadpoolctl>=2.0.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn>=0.23.0->scikit-learn-extra) (3.6.0)\n", "Downloading scikit_learn_extra-0.3.0-cp311-cp311-win_amd64.whl (340 kB)\n", "Installing collected packages: scikit-learn-extra\n", "Successfully installed scikit-learn-extra-0.3.0\n" ] } ], "source": [ "#!pip install scikit-learn-extra" ] }, { "cell_type": "code", "execution_count": 41, "id": "bca0112f", "metadata": {}, "outputs": [], "source": [ "# 分群\n", "from sklearn.cluster import KMeans\n", "from sklearn_extra.cluster import KMedoids\n", "# 只使用word vector 去分群\n", "def cluster(X,method = 'kmeans',n = 2):\n", "\n", " method_dict = {\n", " 'kmeans':KMeans(n_clusters=n, random_state=0),\n", " 'kmedos':KMedoids(n_clusters=n, random_state=0)\n", " }\n", " method_dict[method].fit(X)\n", " result = method_dict[method].predict(X)\n", " return result" ] }, { "cell_type": "code", "execution_count": 42, "id": "8fa7ba74", "metadata": {}, "outputs": [ { "data": { "application/vnd.plotly.v1+json": { "config": { "plotlyServerURL": "https://plot.ly" }, "data": [ { "hovertemplate": "dim1=%{x}
dim2=%{y}
dim3=%{z}
size=%{marker.size}
word=%{text}
color=%{marker.color}", "legendgroup": "", "marker": { "color": { "bdata": "AQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAAAAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAAAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAACAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAADAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAAAQAAAAEAAAABAAAA", "dtype": "i4" }, "coloraxis": "coloraxis", "size": { "bdata": "AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/AAAAAAAA4D8AAAAAAADgPwAAAAAAAOA/", "dtype": "f8" }, "sizemode": "area", "sizeref": 0.00125, "symbol": "circle" }, "mode": "markers+text", "name": "", "scene": "scene", "showlegend": false, "text": [ "抱抱", "扭曲", "禮拜", "實際上", "扣掉", "例子", "男友", "證明", "好好", "消失", "可能性", "打算", "碰到", "分配", "提早", "諮詢", "相關", "照片", "專業", "我能", "更是", "老公", "依舊", "不在乎", "吞下去", "勇敢", "體力", "那位", "墮胎", "假日", "本質", "對話", "法官", "電動", "單方面", "支撐", "這不", "同事", "努力", "講好", "在乎", "項目", "婚姻生活", "居住", "婆媳", "工廠", "生完", "分手", "互動", "要生", "律師", "敘述", "賣掉", "沒差", "停止", "安全感", "盡力", "繼承", "透天", "盡量", "生產", "三次", "陌生人", "老二", "沒事", "來回", "心疼", "發現", "安慰", "所有人", "留言", "表面", "道歉", "接近", "扶養", "頭期", "謝謝", "上次", "第一個", "偷吃", "貸款", "大小事", "男人", "有用", "身邊", "男生", "反駁", "阿公", "無限", "背叛", "困難", "分析", "做出", "丈夫", "未婚夫", "奶瓶", "脾氣", "義務", "你好", "無所謂", "花費", "心情", "做好", "情境", "這份", "拿到", "電視", "養育", "北京", "參與", "扶持", "人才", "犧牲", "失望", "財產", "一路", "觀念", "最終", "危險", "虐待", "怨恨", "改善", "釐清", "磨合", "言論", "情緒", "部份", "我並", "女同事", "提供", "感謝", "贍養費", "打開", "一步", "婚姻", "做法", "姐姐", "言語", "分擔", "長時間", "兩人", "作法", "當你", "拿出", "走路", "撫養費", "伴侶", "管理", "睡眠", "住家" ], "type": "scatter3d", "x": { "bdata": "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", "dtype": "f4" }, "y": { "bdata": "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", "dtype": "f4" }, "z": { "bdata": "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", "dtype": "f4" } } ], "layout": { "coloraxis": { "colorbar": { "title": { "text": "color" } }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ] }, "height": 800, "legend": { "itemsizing": "constant", "tracegroupgap": 0 }, "margin": { "t": 60 }, "scene": { "domain": { "x": [ 0, 1 ], "y": [ 0, 1 ] }, "xaxis": { "title": { "text": "dim1" } }, "yaxis": { "title": { "text": "dim2" } }, "zaxis": { "title": { "text": "dim3" } } }, "template": { "data": { "bar": [ { "error_x": { "color": "#2a3f5f" }, "error_y": { "color": "#2a3f5f" }, "marker": { "line": { "color": "#E5ECF6", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "bar" } ], "barpolar": [ { "marker": { "line": { "color": "#E5ECF6", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "barpolar" } ], "carpet": [ { "aaxis": { "endlinecolor": "#2a3f5f", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "#2a3f5f" }, "baxis": { "endlinecolor": "#2a3f5f", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "#2a3f5f" }, "type": "carpet" } ], "choropleth": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "choropleth" } ], "contour": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "contour" } ], "contourcarpet": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "contourcarpet" } ], "heatmap": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "heatmap" } ], "histogram": [ { "marker": { "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "histogram" } ], "histogram2d": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "histogram2d" } ], "histogram2dcontour": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "histogram2dcontour" } ], "mesh3d": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "type": "mesh3d" } ], "parcoords": [ { "line": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "parcoords" } ], "pie": [ { "automargin": true, "type": "pie" } ], "scatter": [ { "fillpattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 }, "type": "scatter" } ], "scatter3d": [ { "line": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatter3d" } ], "scattercarpet": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattercarpet" } ], "scattergeo": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattergeo" } ], "scattergl": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattergl" } ], "scattermap": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattermap" } ], "scattermapbox": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scattermapbox" } ], "scatterpolar": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterpolar" } ], "scatterpolargl": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterpolargl" } ], "scatterternary": [ { "marker": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "type": "scatterternary" } ], "surface": [ { "colorbar": { "outlinewidth": 0, "ticks": "" }, "colorscale": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "type": "surface" } ], "table": [ { "cells": { "fill": { "color": "#EBF0F8" }, "line": { "color": "white" } }, "header": { "fill": { "color": "#C8D4E3" }, "line": { "color": "white" } }, "type": "table" } ] }, "layout": { "annotationdefaults": { "arrowcolor": "#2a3f5f", "arrowhead": 0, "arrowwidth": 1 }, "autotypenumbers": "strict", "coloraxis": { "colorbar": { "outlinewidth": 0, "ticks": "" } }, "colorscale": { "diverging": [ [ 0, "#8e0152" ], [ 0.1, "#c51b7d" ], [ 0.2, "#de77ae" ], [ 0.3, "#f1b6da" ], [ 0.4, "#fde0ef" ], [ 0.5, "#f7f7f7" ], [ 0.6, "#e6f5d0" ], [ 0.7, "#b8e186" ], [ 0.8, "#7fbc41" ], [ 0.9, "#4d9221" ], [ 1, "#276419" ] ], "sequential": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ], "sequentialminus": [ [ 0, "#0d0887" ], [ 0.1111111111111111, "#46039f" ], [ 0.2222222222222222, "#7201a8" ], [ 0.3333333333333333, "#9c179e" ], [ 0.4444444444444444, "#bd3786" ], [ 0.5555555555555556, "#d8576b" ], [ 0.6666666666666666, "#ed7953" ], [ 0.7777777777777778, "#fb9f3a" ], [ 0.8888888888888888, "#fdca26" ], [ 1, "#f0f921" ] ] }, "colorway": [ "#636efa", "#EF553B", "#00cc96", "#ab63fa", "#FFA15A", "#19d3f3", "#FF6692", "#B6E880", "#FF97FF", "#FECB52" ], "font": { "color": "#2a3f5f" }, "geo": { "bgcolor": "white", "lakecolor": "white", "landcolor": "#E5ECF6", "showlakes": true, "showland": true, "subunitcolor": "white" }, "hoverlabel": { "align": "left" }, "hovermode": "closest", "mapbox": { "style": "light" }, "paper_bgcolor": "white", "plot_bgcolor": "#E5ECF6", "polar": { "angularaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "bgcolor": "#E5ECF6", "radialaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" } }, "scene": { "xaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" }, "yaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" }, "zaxis": { "backgroundcolor": "#E5ECF6", "gridcolor": "white", "gridwidth": 2, "linecolor": "white", "showbackground": true, "ticks": "", "zerolinecolor": "white" } }, "shapedefaults": { "line": { "color": "#2a3f5f" } }, "ternary": { "aaxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "baxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" }, "bgcolor": "#E5ECF6", "caxis": { "gridcolor": "white", "linecolor": "white", "ticks": "" } }, "title": { "x": 0.05 }, "xaxis": { "automargin": true, "gridcolor": "white", "linecolor": "white", "ticks": "", "title": { "standoff": 15 }, "zerolinecolor": "white", "zerolinewidth": 2 }, "yaxis": { "automargin": true, "gridcolor": "white", "linecolor": "white", "ticks": "", "title": { "standoff": 15 }, "zerolinecolor": "white", "zerolinewidth": 2 } } }, "width": 800 } } }, "metadata": {}, "output_type": "display_data" } ], "source": [ "new_feat = reduceDim(feat,method='PCA',dim = 20)\n", "d3_feat = reduceDim(feat,method='PCA',dim = 3)\n", "word_df = pd.DataFrame({\n", " \"word\":sample_words,\n", " \"color\":cluster(new_feat,n=4),\n", " \"dim1\":d3_feat[:,0],\n", " \"dim2\":d3_feat[:,1],\n", " \"dim3\":d3_feat[:,2],\n", "\n", "})\n", "plotScatter3D(word_df)" ] }, { "cell_type": "markdown", "id": "5ac5d686", "metadata": {}, "source": [ "### Transformers Embeddings" ] }, { "cell_type": "markdown", "id": "c2e45969", "metadata": {}, "source": [ "#### 使用 Sentence-Transformer 套件 \n", "參考資料：https://www.sbert.net/index.html" ] }, { "cell_type": "code", "execution_count": null, "id": "244afb48", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Collecting sentence-transformers\n", " Downloading sentence_transformers-4.1.0-py3-none-any.whl.metadata (13 kB)\n", "Collecting transformers<5.0.0,>=4.41.0 (from sentence-transformers)\n", " Downloading transformers-4.51.3-py3-none-any.whl.metadata (38 kB)\n", "Requirement already satisfied: tqdm in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (4.67.1)\n", "Requirement already satisfied: torch>=1.11.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (2.4.1)\n", "Requirement already satisfied: scikit-learn in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (1.4.0)\n", "Requirement already satisfied: scipy in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (1.12.0)\n", "Requirement already satisfied: huggingface-hub>=0.20.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (0.25.1)\n", "Requirement already satisfied: Pillow in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (11.1.0)\n", "Requirement already satisfied: typing_extensions>=4.5.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sentence-transformers) (4.12.2)\n", "Requirement already satisfied: filelock in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from huggingface-hub>=0.20.0->sentence-transformers) (3.18.0)\n", "Requirement already satisfied: fsspec>=2023.5.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from huggingface-hub>=0.20.0->sentence-transformers) (2025.3.2)\n", "Requirement already satisfied: packaging>=20.9 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from huggingface-hub>=0.20.0->sentence-transformers) (24.2)\n", "Requirement already satisfied: pyyaml>=5.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from huggingface-hub>=0.20.0->sentence-transformers) (6.0.2)\n", "Requirement already satisfied: requests in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from huggingface-hub>=0.20.0->sentence-transformers) (2.32.3)\n", "Requirement already satisfied: sympy in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from torch>=1.11.0->sentence-transformers) (1.13.3)\n", "Requirement already satisfied: networkx in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from torch>=1.11.0->sentence-transformers) (3.4.2)\n", "Requirement already satisfied: jinja2 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from torch>=1.11.0->sentence-transformers) (3.1.6)\n", "Requirement already satisfied: colorama in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from tqdm->sentence-transformers) (0.4.6)\n", "Collecting huggingface-hub>=0.20.0 (from sentence-transformers)\n", " Downloading huggingface_hub-0.31.1-py3-none-any.whl.metadata (13 kB)\n", "Requirement already satisfied: numpy>=1.17 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (1.24.3)\n", "Requirement already satisfied: regex!=2019.12.17 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from transformers<5.0.0,>=4.41.0->sentence-transformers) (2024.11.6)\n", "Collecting tokenizers<0.22,>=0.21 (from transformers<5.0.0,>=4.41.0->sentence-transformers)\n", " Downloading tokenizers-0.21.1-cp39-abi3-win_amd64.whl.metadata (6.9 kB)\n", "Collecting safetensors>=0.4.3 (from transformers<5.0.0,>=4.41.0->sentence-transformers)\n", " Downloading safetensors-0.5.3-cp38-abi3-win_amd64.whl.metadata (3.9 kB)\n", "Requirement already satisfied: joblib>=1.2.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn->sentence-transformers) (1.4.2)\n", "Requirement already satisfied: threadpoolctl>=2.0.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from scikit-learn->sentence-transformers) (3.6.0)\n", "Requirement already satisfied: MarkupSafe>=2.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from jinja2->torch>=1.11.0->sentence-transformers) (3.0.2)\n", "Requirement already satisfied: charset-normalizer<4,>=2 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (3.3.2)\n", "Requirement already satisfied: idna<4,>=2.5 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (3.7)\n", "Requirement already satisfied: urllib3<3,>=1.21.1 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (2.3.0)\n", "Requirement already satisfied: certifi>=2017.4.17 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from requests->huggingface-hub>=0.20.0->sentence-transformers) (2025.1.31)\n", "Requirement already satisfied: mpmath<1.4,>=1.1.0 in c:\\users\\rolya\\anaconda3\\envs\\syllabus\\lib\\site-packages (from sympy->torch>=1.11.0->sentence-transformers) (1.3.0)\n", "Downloading sentence_transformers-4.1.0-py3-none-any.whl (345 kB)\n", "Downloading transformers-4.51.3-py3-none-any.whl (10.4 MB)\n", " ---------------------------------------- 0.0/10.4 MB ? eta -:--:--\n", " - -------------------------------------- 0.3/10.4 MB ? eta -:--:--\n", " --- ------------------------------------ 0.8/10.4 MB 2.6 MB/s eta 0:00:04\n", " ------ --------------------------------- 1.6/10.4 MB 3.0 MB/s eta 0:00:03\n", " ------- -------------------------------- 1.8/10.4 MB 2.5 MB/s eta 0:00:04\n", " ---------- ----------------------------- 2.6/10.4 MB 2.7 MB/s eta 0:00:03\n", " ------------ --------------------------- 3.1/10.4 MB 2.8 MB/s eta 0:00:03\n", " --------------- ------------------------ 3.9/10.4 MB 2.9 MB/s eta 0:00:03\n", " ------------------ --------------------- 4.7/10.4 MB 2.9 MB/s eta 0:00:02\n", " -------------------- ------------------- 5.2/10.4 MB 3.0 MB/s eta 0:00:02\n", " ------------------------ --------------- 6.3/10.4 MB 3.1 MB/s eta 0:00:02\n", " --------------------------- ------------ 7.1/10.4 MB 3.2 MB/s eta 0:00:02\n", " ----------------------------- ---------- 7.6/10.4 MB 3.2 MB/s eta 0:00:01\n", " ------------------------------- -------- 8.1/10.4 MB 3.1 MB/s eta 0:00:01\n", " --------------------------------- ------ 8.7/10.4 MB 3.1 MB/s eta 0:00:01\n", " ------------------------------------- -- 9.7/10.4 MB 3.2 MB/s eta 0:00:01\n", " ---------------------------------------- 10.4/10.4 MB 3.2 MB/s eta 0:00:00\n", "Downloading huggingface_hub-0.31.1-py3-none-any.whl (484 kB)\n", "Downloading safetensors-0.5.3-cp38-abi3-win_amd64.whl (308 kB)\n", "Downloading tokenizers-0.21.1-cp39-abi3-win_amd64.whl (2.4 MB)\n", " ---------------------------------------- 0.0/2.4 MB ? eta -:--:--\n", " ---- ----------------------------------- 0.3/2.4 MB ? eta -:--:--\n", " -------- ------------------------------- 0.5/2.4 MB 989.2 kB/s eta 0:00:02\n", " ------------ --------------------------- 0.8/2.4 MB 1.4 MB/s eta 0:00:02\n", " ----------------- ---------------------- 1.0/2.4 MB 1.3 MB/s eta 0:00:02\n", " ------------------------- -------------- 1.6/2.4 MB 1.5 MB/s eta 0:00:01\n", " ---------------------------------- ----- 2.1/2.4 MB 1.7 MB/s eta 0:00:01\n", " ---------------------------------------- 2.4/2.4 MB 1.7 MB/s eta 0:00:00\n", "Installing collected packages: safetensors, huggingface-hub, tokenizers, transformers, sentence-transformers\n", " Attempting uninstall: huggingface-hub\n", " Found existing installation: huggingface-hub 0.25.1\n", " Uninstalling huggingface-hub-0.25.1:\n", " Successfully uninstalled huggingface-hub-0.25.1\n", "Successfully installed huggingface-hub-0.31.1 safetensors-0.5.3 sentence-transformers-4.1.0 tokenizers-0.21.1 transformers-4.51.3\n" ] } ], "source": [ "#!pip install -U sentence-transformers" ] }, { "cell_type": "code", "execution_count": 50, "id": "9fd5635e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "WARNING:tensorflow:From c:\\Users\\rolya\\anaconda3\\envs\\Syllabus\\Lib\\site-packages\\keras\\src\\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.\n", "\n" ] } ], "source": [ "from sentence_transformers import SentenceTransformer, models, util" ] }, { "cell_type": "markdown", "id": "fe816241", "metadata": {}, "source": [ "#### 小模型，以BERT為範例" ] }, { "cell_type": "markdown", "id": "473a459b", "metadata": {}, "source": [ "中文 bert-base-chinese" ] }, { "cell_type": "code", "execution_count": null, "id": "3fb7ca1c", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "No sentence-transformers model found with name google-bert/bert-base-chinese. Creating a new one with mean pooling.\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "1e010457b0b44786a271e7a2c7032dbe", "version_major": 2, "version_minor": 0 }, "text/plain": [ "config.json: 0%| | 0.00/624 [00:00\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
sentence1sentence2score
0今天天氣很好。今天是個晴空萬里的好天氣。0.914067
1今天天氣很好。我晚上想去公園散步。0.787589
2今天是個晴空萬里的好天氣。我晚上想去公園散步。0.759642
\n", "" ], "text/plain": [ " sentence1 sentence2 score\n", "0 今天天氣很好。 今天是個晴空萬里的好天氣。 0.914067\n", "1 今天天氣很好。 我晚上想去公園散步。 0.787589\n", "2 今天是個晴空萬里的好天氣。 我晚上想去公園散步。 0.759642" ] }, "execution_count": 53, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 示範句子\n", "sentences = [\n", " \"今天天氣很好。\",\n", " \"今天是個晴空萬里的好天氣。\",\n", " \"我晚上想去公園散步。\"\n", "]\n", "\n", "# 使用 encode() 對資料做embedding\n", "embeddings_ch = bert_ch.encode(sentences)\n", "\n", "# Compute cosine-similarities\n", "cosine_scores = util.cos_sim(embeddings_ch, embeddings_ch)\n", "\n", "# 印出句子間的cosine similarity分數\n", "result = []\n", "for i in range(len(sentences)):\n", " for j in range(i+1, len(sentences)):\n", " result.append([sentences[i], sentences[j], cosine_scores[i][j].item()])\n", "\n", "result_df = pd.DataFrame(result, columns=[\"sentence1\", \"sentence2\", \"score\"])\n", "result_df.sort_values(\"score\", ascending = False)" ] }, { "cell_type": "markdown", "id": "3417de4d", "metadata": {}, "source": [ "### 使用embedding做NLP任務" ] }, { "cell_type": "markdown", "id": "24e475d1", "metadata": {}, "source": [ "#### 相似文件" ] }, { "cell_type": "code", "execution_count": 58, "id": "20e70d68", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "C:\\Users\\rolya\\AppData\\Local\\Temp\\ipykernel_13724\\852741673.py:2: SettingWithCopyWarning:\n", "\n", "\n", "A value is trying to be set on a copy of a slice from a DataFrame.\n", "Try using .loc[row_indexer,col_indexer] = value instead\n", "\n", "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", "\n" ] }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartTitleartContent
01[求助]真的可以不給看手機嗎？常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒絕嗎感覺你拒絕就是心裡有...
12Re:老公工作不穩定因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到這世間只是苦難...
23Re:[求助]真的可以不給看手機嗎？手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎三隻四隻不夠可以辦五隻十...
34[心情]我搞不懂老公到底在想甚麼其實都是小事但都可以吵到離婚可能我們就是幾歲的小孩昨晚上床睡覺後一直覺得很冷老公也喊冷想說睡...
45Re:[心情]我搞不懂老公到底在想甚麼把棉被翻好正面嗯嗯有嗯那你幹嘛抱怨你老公不是誰上床睡覺還會檢查棉被正反的嗎我我也覺得你半夜叫...
\n", "
" ], "text/plain": [ " system_id artTitle \\\n", "0 1 [求助]真的可以不給看手機嗎？ \n", "1 2 Re:老公工作不穩定 \n", "2 3 Re:[求助]真的可以不給看手機嗎？ \n", "3 4 [心情]我搞不懂老公到底在想甚麼 \n", "4 5 Re:[心情]我搞不懂老公到底在想甚麼 \n", "\n", " artContent \n", "0 常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒絕嗎感覺你拒絕就是心裡有... \n", "1 因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到這世間只是苦難... \n", "2 手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎三隻四隻不夠可以辦五隻十... \n", "3 其實都是小事但都可以吵到離婚可能我們就是幾歲的小孩昨晚上床睡覺後一直覺得很冷老公也喊冷想說睡... \n", "4 把棉被翻好正面嗯嗯有嗯那你幹嘛抱怨你老公不是誰上床睡覺還會檢查棉被正反的嗎我我也覺得你半夜叫... " ] }, "execution_count": 58, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df_similar = origin_data[['system_id','artTitle', 'artContent']]\n", "df_similar['artContent'] = df_similar['artContent'].apply(lambda x: re.sub('[^\\u4e00-\\u9fff]+', '',x))\n", "\n", "df_similar.head(5)" ] }, { "cell_type": "markdown", "id": "0d5fac26", "metadata": {}, "source": [ "使用 bert-base-chinese 做示範" ] }, { "cell_type": "markdown", "id": "e91d233d", "metadata": {}, "source": [ "取得整個文集的 embeddings" ] }, { "cell_type": "code", "execution_count": 60, "id": "fc238105", "metadata": {}, "outputs": [], "source": [ "corpus_embeddings = bert_ch.encode(\n", " df_similar['artContent'],\n", " convert_to_tensor=True,\n", " batch_size=32\n", ")" ] }, { "cell_type": "code", "execution_count": 61, "id": "3781570f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "\n", "======================\n", "\n", "\n", "Query: Re:[閒聊]離婚的掙扎\n", "\n", " 資料集中前五相似的文章:\n", "Re:[閒聊]離婚的掙扎 (Score: 1.0000)\n", "Re:兩人的溝通與目前的情況 (Score: 0.9643)\n", "Re:[心情]人生再重來一次，我不會生小孩 (Score: 0.9635)\n", "Re:[求助]先生的女助理 (Score: 0.9628)\n", "Re:[心情]老公喝酒不懂克制 (Score: 0.9627)\n", "\n", "\n", "======================\n", "\n", "\n" ] } ], "source": [ "query_num = 6 # 指定文章\n", "\n", "# Find the closest 5 sentences of the corpus for each query sentence based on cosine similarity\n", "top_k = 5\n", "\n", "\n", "query_embedding = bert_ch.encode(df_similar['artContent'][query_num], convert_to_tensor=True)\n", "\n", "# We use cosine-similarity and torch.topk to find the highest 5 scores\n", "cos_scores = util.cos_sim(query_embedding, corpus_embeddings)[0]\n", "top_results = torch.topk(cos_scores, k=top_k)\n", "\n", "print(\"\\n\\n======================\\n\\n\")\n", "print(\"Query:\", df_similar['artTitle'][query_num])\n", "print(\"\\n 資料集中前五相似的文章:\")\n", "\n", "for score, idx in zip(top_results[0], top_results[1]):\n", " print(df_similar['artTitle'][idx.item()], \"(Score: {:.4f})\".format(score))\n", "\n", "print(\"\\n\\n======================\\n\\n\")" ] }, { "cell_type": "code", "execution_count": 62, "id": "1ff9051d", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", "\n", "======================\n", "\n", "\n", "Query: [閒聊]美女朋友婚後一直找人約砲\n", "\n", " 資料集中前五相似的文章:\n", "[閒聊]美女朋友婚後一直找人約砲 (Score: 1.0000)\n", "該繼續挽回，還是該放手了？ (Score: 0.9864)\n", "[求助]精神出軌後成功修復感情經驗? (Score: 0.9856)\n", "[閒聊]人妻外食被搞懷孕後，繼續外食？ (Score: 0.9849)\n", "Re:[閒聊]老公有異性友人 (Score: 0.9847)\n", "\n", "\n", "======================\n", "\n", "\n" ] } ], "source": [ "query_num = 30\n", "\n", "top_k = 5\n", "\n", "query_embedding = bert_ch.encode(df_similar['artContent'][query_num], convert_to_tensor=True)\n", "\n", "# We use cosine-similarity and torch.topk to find the highest 5 scores\n", "cos_scores = util.cos_sim(query_embedding, corpus_embeddings)[0]\n", "top_results = torch.topk(cos_scores, k=top_k)\n", "\n", "print(\"\\n\\n======================\\n\\n\")\n", "print(\"Query:\", df_similar['artTitle'][query_num])\n", "print(\"\\n 資料集中前五相似的文章:\")\n", "\n", "for score, idx in zip(top_results[0], top_results[1]):\n", " print(df_similar['artTitle'][idx.item()], \"(Score: {:.4f})\".format(score))\n", "\n", "print(\"\\n\\n======================\\n\\n\")" ] }, { "cell_type": "markdown", "id": "00bb07f0", "metadata": {}, "source": [ "### 分類任務\n", "使用bert-base-chinese模型對ppt文集做embeddings，接著訓練分類器。（參考week7程式碼）" ] }, { "cell_type": "code", "execution_count": 63, "id": "c2e04f29", "metadata": {}, "outputs": [], "source": [ "from sklearn.model_selection import train_test_split\n", "from sklearn.linear_model import LogisticRegression" ] }, { "cell_type": "code", "execution_count": 64, "id": "75cdf073", "metadata": {}, "outputs": [], "source": [ "from sentence_transformers import SentenceTransformer, models, util" ] }, { "cell_type": "markdown", "id": "fe581be5", "metadata": {}, "source": [ "- 版別：婚姻版 結婚版 八卦版\n", "- 工作流程平台搜尋/排除關鍵字：\n", "\n", "文章數分別為
\n", "Gossiping 3380
\n", "GetMarry 2531
\n", "marriage 2204" ] }, { "cell_type": "code", "execution_count": 71, "id": "17045635", "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "\n", "# 讀取三個 CSV 檔案\n", "df1 = pd.read_csv('/Users/rolya/Desktop/DIV/divoce_project2/text.csv')\n", "df2 = pd.read_csv('/Users/rolya/Desktop/DIV/divoce_project2/GetMarry.csv')\n", "df3 = pd.read_csv('/Users/rolya/Desktop/DIV/divoce_project2/Gossiping.csv')\n", "\n", "# 合併成一個 DataFrame\n", "merged_df = pd.concat([df1, df2, df3], ignore_index=True)\n", "\n", "# 儲存為新的 merge.csv\n", "merged_df.to_csv('merge.csv', index=False)\n" ] }, { "cell_type": "code", "execution_count": null, "id": "981028f4", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartUrlartTitleartDateartPosterartCatagoryartContentartCommente_ipinsertedDatedataSource
01https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05safelovemarriage常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"FlyOncidium\"...114.137.169.1052021-01-10 01:20:56ptt
12https://www.ptt.cc/bbs/marriage/M.1610162736.A...Re:老公工作不穩定2021-01-09 11:25:34maykomarriage因為男方工作不穩定,房東才不肯租,只好換女生承租,對吧\\n很殘忍的說...這無緣的孩子聰明,...[]36.229.84.2292021-01-10 01:20:56ptt
23https://www.ptt.cc/bbs/marriage/M.1610190309.A...Re:[求助]真的可以不給看手機嗎？2021-01-09 19:05:00loser1marriage手機要看就給看啊！\\n先帝爺不是說，\\n一隻不夠，不能辦兩隻嗎？\\n兩隻不夠，不能辦三隻嗎？...[{\"cmtStatus\": \"噓\", \"cmtPoster\": \"mark0204\", \"...118.170.238.1382021-01-10 01:20:56ptt
\n", "
" ], "text/plain": [ " system_id artUrl \\\n", "0 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "1 2 https://www.ptt.cc/bbs/marriage/M.1610162736.A... \n", "2 3 https://www.ptt.cc/bbs/marriage/M.1610190309.A... \n", "\n", " artTitle artDate artPoster artCatagory \\\n", "0 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 safelove marriage \n", "1 Re:老公工作不穩定 2021-01-09 11:25:34 mayko marriage \n", "2 Re:[求助]真的可以不給看手機嗎？ 2021-01-09 19:05:00 loser1 marriage \n", "\n", " artContent \\\n", "0 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... \n", "1 因為男方工作不穩定,房東才不肯租,只好換女生承租,對吧\\n很殘忍的說...這無緣的孩子聰明,... \n", "2 手機要看就給看啊！\\n先帝爺不是說，\\n一隻不夠，不能辦兩隻嗎？\\n兩隻不夠，不能辦三隻嗎？... \n", "\n", " artComment e_ip \\\n", "0 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"FlyOncidium\"... 114.137.169.105 \n", "1 [] 36.229.84.229 \n", "2 [{\"cmtStatus\": \"噓\", \"cmtPoster\": \"mark0204\", \"... 118.170.238.138 \n", "\n", " insertedDate dataSource \n", "0 2021-01-10 01:20:56 ptt \n", "1 2021-01-10 01:20:56 ptt \n", "2 2021-01-10 01:20:56 ptt " ] }, "execution_count": 72, "metadata": {}, "output_type": "execute_result" } ], "source": [ "div = pd.read_csv(\"merge.csv\")\n", "div.head(3)" ] }, { "cell_type": "markdown", "id": "7dec8c37", "metadata": {}, "source": [ "## 資料清理\n", "利用標點符號斷句\n", "\n", "文集的標題和內容納入分析的內容，成為content欄位" ] }, { "cell_type": "code", "execution_count": null, "id": "f566c38c", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
contentartUrlartCatagory
0求助真的可以不給看手機嗎常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒...https://www.ptt.cc/bbs/marriage/M.1610159827.A...marriage
1老公工作不穩定因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到...https://www.ptt.cc/bbs/marriage/M.1610162736.A...marriage
2求助真的可以不給看手機嗎手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎...https://www.ptt.cc/bbs/marriage/M.1610190309.A...marriage
3心情我搞不懂老公到底在想甚麼其實都是小事但都可以吵到離婚可能我們就是幾歲的小孩昨晚上床睡覺後...https://www.ptt.cc/bbs/marriage/M.1610193770.A...marriage
4心情我搞不懂老公到底在想甚麼把棉被翻好正面嗯嗯有嗯那你幹嘛抱怨你老公不是誰上床睡覺還會檢查棉...https://www.ptt.cc/bbs/marriage/M.1610203445.A...marriage
\n", "
" ], "text/plain": [ " content \\\n", "0 求助真的可以不給看手機嗎常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒... \n", "1 老公工作不穩定因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到... \n", "2 求助真的可以不給看手機嗎手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎... \n", "3 心情我搞不懂老公到底在想甚麼其實都是小事但都可以吵到離婚可能我們就是幾歲的小孩昨晚上床睡覺後... \n", "4 心情我搞不懂老公到底在想甚麼把棉被翻好正面嗯嗯有嗯那你幹嘛抱怨你老公不是誰上床睡覺還會檢查棉... \n", "\n", " artUrl artCatagory \n", "0 https://www.ptt.cc/bbs/marriage/M.1610159827.A... marriage \n", "1 https://www.ptt.cc/bbs/marriage/M.1610162736.A... marriage \n", "2 https://www.ptt.cc/bbs/marriage/M.1610190309.A... marriage \n", "3 https://www.ptt.cc/bbs/marriage/M.1610193770.A... marriage \n", "4 https://www.ptt.cc/bbs/marriage/M.1610203445.A... marriage " ] }, "execution_count": 73, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 過濾 nan 的資料\n", "div = div.dropna(subset=['artTitle'])\n", "div = div.dropna(subset=['artContent'])\n", "# 移除網址格式\n", "div[\"artContent\"] = div.artContent.apply(\n", " lambda x: re.sub(\"(http|https)://.*\", \"\", x)\n", ")\n", "div[\"artTitle\"] = div[\"artTitle\"].apply(\n", " lambda x: re.sub(\"(http|https)://.*\", \"\", x)\n", ")\n", "# 只留下中文字\n", "div[\"artContent\"] = div.artContent.apply(\n", " lambda x: re.sub(\"[^\\u4e00-\\u9fa5]+\", \"\", x)\n", ")\n", "div[\"artTitle\"] = div[\"artTitle\"].apply(\n", " lambda x: re.sub(\"[^\\u4e00-\\u9fa5]+\", \"\", x)\n", ")\n", "\n", "# 留下 content\n", "div[\"content\"] = div[\"artTitle\"] + div[\"artContent\"]\n", "div = div[[\"content\", \"artUrl\", \"artCatagory\"]] # 文章內容 文章連結\n", "div.head()" ] }, { "cell_type": "markdown", "id": "5b0a2414", "metadata": {}, "source": [ "### 使用Bert做encoding" ] }, { "cell_type": "code", "execution_count": null, "id": "78700878", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
contentartUrlartCatagoryembeddings
0求助真的可以不給看手機嗎常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒...https://www.ptt.cc/bbs/marriage/M.1610159827.A...marriage[0.8504211, -0.23771943, -0.14535092, 0.156168...
1老公工作不穩定因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到...https://www.ptt.cc/bbs/marriage/M.1610162736.A...marriage[0.7396465, -0.14757201, -0.057030175, 0.26861...
2求助真的可以不給看手機嗎手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎...https://www.ptt.cc/bbs/marriage/M.1610190309.A...marriage[0.4144313, -0.095341206, -0.2705029, 0.380942...
\n", "
" ], "text/plain": [ " content \\\n", "0 求助真的可以不給看手機嗎常常看大家說偷看手機是不對的但如果已經結婚了太太想看你手機真的可以拒... \n", "1 老公工作不穩定因為男方工作不穩定房東才不肯租只好換女生承租對吧很殘忍的說這無緣的孩子聰明來到... \n", "2 求助真的可以不給看手機嗎手機要看就給看啊先帝爺不是說一隻不夠不能辦兩隻嗎兩隻不夠不能辦三隻嗎... \n", "\n", " artUrl artCatagory \\\n", "0 https://www.ptt.cc/bbs/marriage/M.1610159827.A... marriage \n", "1 https://www.ptt.cc/bbs/marriage/M.1610162736.A... marriage \n", "2 https://www.ptt.cc/bbs/marriage/M.1610190309.A... marriage \n", "\n", " embeddings \n", "0 [0.8504211, -0.23771943, -0.14535092, 0.156168... \n", "1 [0.7396465, -0.14757201, -0.057030175, 0.26861... \n", "2 [0.4144313, -0.095341206, -0.2705029, 0.380942... " ] }, "execution_count": 74, "metadata": {}, "output_type": "execute_result" } ], "source": [ "div[\"embeddings\"] = div.content.apply(lambda x: bert_ch.encode(x))\n", "div.head(3)" ] }, { "cell_type": "code", "execution_count": 75, "id": "48df0ffb", "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "from ast import literal_eval" ] }, { "cell_type": "code", "execution_count": null, "id": "51a5d912", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " 0 1 2 3 4 5 6 \\\n", "5325 0.578834 -0.050640 -0.503452 0.114320 0.008045 0.006949 -0.056759 \n", "6304 0.393490 0.062839 -0.422073 0.157619 0.078032 -0.188103 0.018693 \n", "1659 0.629233 -0.231260 -0.266918 0.320200 -0.087541 -0.173598 -0.265297 \n", "1280 0.670054 -0.171460 -0.248393 0.444933 -0.170084 -0.130054 -0.090347 \n", "291 0.621767 -0.093077 -0.131523 0.232634 -0.262020 -0.321169 -0.187148 \n", "\n", " 7 8 9 ... 758 759 760 \\\n", "5325 0.107957 -0.071790 -0.321657 ... -0.201785 -0.409468 0.127512 \n", "6304 0.168009 -0.057865 -0.318298 ... -0.059410 -0.456915 0.209086 \n", "1659 0.012857 -0.264234 -0.477913 ... 0.013435 -0.425337 0.112885 \n", "1280 -0.076344 -0.472568 -0.343588 ... -0.273166 -0.459275 0.273230 \n", "291 -0.086306 -0.223639 -0.403096 ... -0.066033 -0.425651 0.224642 \n", "\n", " 761 762 763 764 765 766 767 \n", "5325 0.211407 -0.226111 0.095762 -0.286687 0.348344 0.171068 0.239167 \n", "6304 -0.012599 -0.240362 0.116647 -0.470287 0.465576 0.112342 0.218261 \n", "1659 0.025797 -0.122397 0.304335 -0.274302 0.218617 0.328430 0.069736 \n", "1280 -0.121515 0.074967 0.099266 -0.279123 0.051970 0.352049 -0.089615 \n", "291 -0.060458 -0.125518 0.139101 -0.270676 0.286504 0.216595 0.045877 \n", "\n", "[5 rows x 768 columns]\n", "5325 Gossiping\n", "6304 Gossiping\n", "1659 marriage\n", "1280 marriage\n", "291 marriage\n", "Name: artCatagory, dtype: object\n" ] } ], "source": [ "data = div.copy()\n", "\n", "X = data[\"embeddings\"].apply(pd.Series)\n", "y = data[\"artCatagory\"]\n", "\n", "# 把整個資料集七三切\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y, test_size=0.3, random_state=777\n", ")\n", "\n", "print(X_train.head())\n", "print(y_train.head())" ] }, { "cell_type": "code", "execution_count": 77, "id": "ef7fcdd6", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "c:\\Users\\rolya\\anaconda3\\envs\\Syllabus\\Lib\\site-packages\\sklearn\\linear_model\\_logistic.py:469: ConvergenceWarning:\n", "\n", "lbfgs failed to converge (status=1):\n", "STOP: TOTAL NO. of ITERATIONS REACHED LIMIT.\n", "\n", "Increase the number of iterations (max_iter) or scale the data as shown in:\n", " https://scikit-learn.org/stable/modules/preprocessing.html\n", "Please also refer to the documentation for alternative solver options:\n", " https://scikit-learn.org/stable/modules/linear_model.html#logistic-regression\n", "\n" ] }, { "data": { "text/html": [ "
LogisticRegression()
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
" ], "text/plain": [ "LogisticRegression()" ] }, "execution_count": 77, "metadata": {}, "output_type": "execute_result" } ], "source": [ "clf = LogisticRegression()\n", "clf.fit(X_train, y_train)\n", "clf" ] }, { "cell_type": "code", "execution_count": 78, "id": "53f8dd9e", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "['Gossiping' 'Gossiping' 'Gossiping' 'marriage' 'GetMarry' 'marriage'\n", " 'marriage' 'GetMarry' 'marriage' 'Gossiping']\n" ] } ], "source": [ "y_pred = clf.predict(X_test)\n", "y_pred_proba = clf.predict_proba(X_test)\n", "print(y_pred[:10])" ] }, { "cell_type": "code", "execution_count": 79, "id": "ef647d78", "metadata": {}, "outputs": [], "source": [ "from sklearn.metrics import classification_report" ] }, { "cell_type": "code", "execution_count": 80, "id": "707f94a0", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", " GetMarry 1.00 1.00 1.00 768\n", " Gossiping 0.99 0.99 0.99 1011\n", " marriage 0.99 0.99 0.99 655\n", "\n", " accuracy 0.99 2434\n", " macro avg 0.99 0.99 0.99 2434\n", "weighted avg 0.99 0.99 0.99 2434\n", "\n" ] } ], "source": [ "## Accuracy, Precision, Recall, F1-score\n", "print(classification_report(y_test, y_pred))" ] }, { "cell_type": "markdown", "id": "bb0e8f1f", "metadata": {}, "source": [ "## 用訓練好的分類器來預測不同時間的文章類別" ] }, { "cell_type": "code", "execution_count": 100, "id": "08776b55", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartUrlartTitleartDateartPosterartCatagoryartContentartCommente_ipinsertedDatedataSource
01https://www.ptt.cc/bbs/GetMarry/M.1653965335.A...[推薦]彰化_DeerHer客製手工喜餅2022-05-31 10:48:52piemeGetMarry廠商所在地區:彰化\\n\\n是什麼場合用到:訂結婚_111/5\\n\\n廠商名稱:\\n喜餅:\\n...[]36.232.149.352022-06-01 01:14:50ptt
12https://www.ptt.cc/bbs/GetMarry/M.1653974818.A...[分享]台北/荳蔻攝影工作室婚紗照2022-05-31 13:26:56ajjhhjGetMarry剛開始查婚紗資訊，真的好討厭傳統婚紗店的組數限制，或是各種無止境加購方案，也好\\n怕遇到纏人...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"exorcist1\", ...223.141.4.2152022-06-01 01:14:50ptt
23https://www.ptt.cc/bbs/GetMarry/M.1654041655.A...[請益]Oohlalove喜餅品項選擇2022-06-01 08:00:53love07erikaGetMarry選擇障礙的新娘來求助了！\\n\\nC區的選擇掙扎到要給品項的deadline…\\n目前確定抹茶...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"mimiwei955\",...111.82.79.252022-06-02 01:14:43ptt
34https://www.ptt.cc/bbs/GetMarry/M.1654042316.A...[贈送]白色及粉色小禮服2022-06-01 08:11:54cherishposseGetMarry《洽中，暫勿來信》\\n\\n已結婚完幾年了才發現還有兩件小禮服擱置在家裡，因家人暫時也用不到了...[]101.10.0.1492022-06-02 01:14:43ptt
45https://www.ptt.cc/bbs/GetMarry/M.1654050290.A...[廣宣]OohLaLove喜餅2022-06-01 10:24:48michael9586GetMarry新人or廠商所在地區:台中\\n\\n是屬於新人哪種場合:結婚 2022/12/18\\n\\n\\n...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"kenkao25\", \"...223.138.172.862022-06-02 01:14:43ptt
....................................
52221375https://www.ptt.cc/bbs/marriage/M.1672325600.A...Re:[心情]果然跟版上說的一樣，還是得離婚...2022-12-29 22:53:18aass5566marriage感謝這個決定？\\n\\n我是覺得根本是被這個諮商害到了吧？\\n\\n如果不是弄這個諮商\\n\\n早...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"wts4832\", \"c...49.158.132.1192022-12-30 01:57:14ptt
52231376https://www.ptt.cc/bbs/marriage/M.1672369383.A...Re:[求助]老公偷拍女同事腿2022-12-30 11:03:00GunOfWindmarriage先說結論 離婚吧\\n沒有小孩 你不能接受 就離婚 +1 吧\\n程度問題\\n就像有些人會...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"robertdelun\"...125.227.145.312022-12-31 01:52:50ptt
52241377https://www.ptt.cc/bbs/marriage/M.1672444694.A...Re:[閒聊]另一半的家庭觀念2022-12-31 07:58:12magicbook123marriage首先，有問題不要上來發文\\n\\n這裡是離婚板 不會給什麼好建議 開口先喊離婚\\n\\n就算有正...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"mtyc\", \"cmtC...223.137.86.712023-01-01 01:51:58ptt
52251378https://www.ptt.cc/bbs/marriage/M.1672455232.A...Re:[閒聊]另一半的家庭觀念2022-12-31 10:53:50lamabclamabcmarriage原PO我有看你舊文，也有留意你的補充和回文裏的推文，看得出來你真的很困擾。\\n\\n很高興你決...[{\"cmtStatus\": \"推\", \"cmtPoster\": \"lastever\", \"...155.137.208.192023-01-01 01:51:58ptt
52261379https://www.ptt.cc/bbs/marriage/M.1672479128.A...[閒聊]離婚，關於結婚金飾2022-12-31 17:32:06penchlinmarriage想請教各位，怎樣比較合理\\n1、男女各拿回自己買的\\n2、男女各拿回送給對方的\\n3、其他\\n[{\"cmtStatus\": \"噓\", \"cmtPoster\": \"wnwe\", \"cmtC...180.217.44.2472023-01-01 01:51:58ptt
\n", "

5156 rows × 11 columns

\n", "
" ], "text/plain": [ " system_id artUrl \\\n", "0 1 https://www.ptt.cc/bbs/GetMarry/M.1653965335.A... \n", "1 2 https://www.ptt.cc/bbs/GetMarry/M.1653974818.A... \n", "2 3 https://www.ptt.cc/bbs/GetMarry/M.1654041655.A... \n", "3 4 https://www.ptt.cc/bbs/GetMarry/M.1654042316.A... \n", "4 5 https://www.ptt.cc/bbs/GetMarry/M.1654050290.A... \n", "... ... ... \n", "5222 1375 https://www.ptt.cc/bbs/marriage/M.1672325600.A... \n", "5223 1376 https://www.ptt.cc/bbs/marriage/M.1672369383.A... \n", "5224 1377 https://www.ptt.cc/bbs/marriage/M.1672444694.A... \n", "5225 1378 https://www.ptt.cc/bbs/marriage/M.1672455232.A... \n", "5226 1379 https://www.ptt.cc/bbs/marriage/M.1672479128.A... \n", "\n", " artTitle artDate artPoster \\\n", "0 [推薦]彰化_DeerHer客製手工喜餅 2022-05-31 10:48:52 pieme \n", "1 [分享]台北/荳蔻攝影工作室婚紗照 2022-05-31 13:26:56 ajjhhj \n", "2 [請益]Oohlalove喜餅品項選擇 2022-06-01 08:00:53 love07erika \n", "3 [贈送]白色及粉色小禮服 2022-06-01 08:11:54 cherishposse \n", "4 [廣宣]OohLaLove喜餅 2022-06-01 10:24:48 michael9586 \n", "... ... ... ... \n", "5222 Re:[心情]果然跟版上說的一樣，還是得離婚... 2022-12-29 22:53:18 aass5566 \n", "5223 Re:[求助]老公偷拍女同事腿 2022-12-30 11:03:00 GunOfWind \n", "5224 Re:[閒聊]另一半的家庭觀念 2022-12-31 07:58:12 magicbook123 \n", "5225 Re:[閒聊]另一半的家庭觀念 2022-12-31 10:53:50 lamabclamabc \n", "5226 [閒聊]離婚，關於結婚金飾 2022-12-31 17:32:06 penchlin \n", "\n", " artCatagory artContent \\\n", "0 GetMarry 廠商所在地區:彰化\\n\\n是什麼場合用到:訂結婚_111/5\\n\\n廠商名稱:\\n喜餅:\\n... \n", "1 GetMarry 剛開始查婚紗資訊，真的好討厭傳統婚紗店的組數限制，或是各種無止境加購方案，也好\\n怕遇到纏人... \n", "2 GetMarry 選擇障礙的新娘來求助了！\\n\\nC區的選擇掙扎到要給品項的deadline…\\n目前確定抹茶... \n", "3 GetMarry 《洽中，暫勿來信》\\n\\n已結婚完幾年了才發現還有兩件小禮服擱置在家裡，因家人暫時也用不到了... \n", "4 GetMarry 新人or廠商所在地區:台中\\n\\n是屬於新人哪種場合:結婚 2022/12/18\\n\\n\\n... \n", "... ... ... \n", "5222 marriage 感謝這個決定？\\n\\n我是覺得根本是被這個諮商害到了吧？\\n\\n如果不是弄這個諮商\\n\\n早... \n", "5223 marriage 先說結論 離婚吧\\n沒有小孩 你不能接受 就離婚 +1 吧\\n程度問題\\n就像有些人會... \n", "5224 marriage 首先，有問題不要上來發文\\n\\n這裡是離婚板 不會給什麼好建議 開口先喊離婚\\n\\n就算有正... \n", "5225 marriage 原PO我有看你舊文，也有留意你的補充和回文裏的推文，看得出來你真的很困擾。\\n\\n很高興你決... \n", "5226 marriage 想請教各位，怎樣比較合理\\n1、男女各拿回自己買的\\n2、男女各拿回送給對方的\\n3、其他\\n \n", "\n", " artComment e_ip \\\n", "0 [] 36.232.149.35 \n", "1 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"exorcist1\", ... 223.141.4.215 \n", "2 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"mimiwei955\",... 111.82.79.25 \n", "3 [] 101.10.0.149 \n", "4 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"kenkao25\", \"... 223.138.172.86 \n", "... ... ... \n", "5222 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"wts4832\", \"c... 49.158.132.119 \n", "5223 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"robertdelun\"... 125.227.145.31 \n", "5224 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"mtyc\", \"cmtC... 223.137.86.71 \n", "5225 [{\"cmtStatus\": \"推\", \"cmtPoster\": \"lastever\", \"... 155.137.208.19 \n", "5226 [{\"cmtStatus\": \"噓\", \"cmtPoster\": \"wnwe\", \"cmtC... 180.217.44.247 \n", "\n", " insertedDate dataSource \n", "0 2022-06-01 01:14:50 ptt \n", "1 2022-06-01 01:14:50 ptt \n", "2 2022-06-02 01:14:43 ptt \n", "3 2022-06-02 01:14:43 ptt \n", "4 2022-06-02 01:14:43 ptt \n", "... ... ... \n", "5222 2022-12-30 01:57:14 ptt \n", "5223 2022-12-31 01:52:50 ptt \n", "5224 2023-01-01 01:51:58 ptt \n", "5225 2023-01-01 01:51:58 ptt \n", "5226 2023-01-01 01:51:58 ptt \n", "\n", "[5156 rows x 11 columns]" ] }, "execution_count": 100, "metadata": {}, "output_type": "execute_result" } ], "source": [ "file_list = [\"GetMarry_t.csv\", \"Gossiping_t.csv\", \"marriage_t.csv\"]\n", "dfs = [pd.read_csv(f) for f in file_list]\n", "ct = pd.concat(dfs, ignore_index=True)\n", "ct.dropna(inplace=True)\n", "ct" ] }, { "cell_type": "code", "execution_count": 101, "id": "e783ecec", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
contentartUrlartCatagory
0推薦彰化客製手工喜餅廠商所在地區彰化是什麼場合用到訂結婚廠商名稱喜餅手工喜餅聯絡資訊官網訂購...https://www.ptt.cc/bbs/GetMarry/M.1653965335.A...GetMarry
1分享台北荳蔻攝影工作室婚紗照剛開始查婚紗資訊真的好討厭傳統婚紗店的組數限制或是各種無止境加購...https://www.ptt.cc/bbs/GetMarry/M.1653974818.A...GetMarry
2請益喜餅品項選擇選擇障礙的新娘來求助了區的選擇掙扎到要給品項的目前確定抹茶鹽之花另一個想要檸...https://www.ptt.cc/bbs/GetMarry/M.1654041655.A...GetMarry
3贈送白色及粉色小禮服洽中暫勿來信已結婚完幾年了才發現還有兩件小禮服擱置在家裡因家人暫時也用不...https://www.ptt.cc/bbs/GetMarry/M.1654042316.A...GetMarry
4廣宣喜餅新人廠商所在地區台中是屬於新人哪種場合結婚以上三項依照要推薦的廠商類別而填寫不得不填...https://www.ptt.cc/bbs/GetMarry/M.1654050290.A...GetMarry
\n", "
" ], "text/plain": [ " content \\\n", "0 推薦彰化客製手工喜餅廠商所在地區彰化是什麼場合用到訂結婚廠商名稱喜餅手工喜餅聯絡資訊官網訂購... \n", "1 分享台北荳蔻攝影工作室婚紗照剛開始查婚紗資訊真的好討厭傳統婚紗店的組數限制或是各種無止境加購... \n", "2 請益喜餅品項選擇選擇障礙的新娘來求助了區的選擇掙扎到要給品項的目前確定抹茶鹽之花另一個想要檸... \n", "3 贈送白色及粉色小禮服洽中暫勿來信已結婚完幾年了才發現還有兩件小禮服擱置在家裡因家人暫時也用不... \n", "4 廣宣喜餅新人廠商所在地區台中是屬於新人哪種場合結婚以上三項依照要推薦的廠商類別而填寫不得不填... \n", "\n", " artUrl artCatagory \n", "0 https://www.ptt.cc/bbs/GetMarry/M.1653965335.A... GetMarry \n", "1 https://www.ptt.cc/bbs/GetMarry/M.1653974818.A... GetMarry \n", "2 https://www.ptt.cc/bbs/GetMarry/M.1654041655.A... GetMarry \n", "3 https://www.ptt.cc/bbs/GetMarry/M.1654042316.A... GetMarry \n", "4 https://www.ptt.cc/bbs/GetMarry/M.1654050290.A... GetMarry " ] }, "execution_count": 101, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 過濾 nan 的資料\n", "ct = ct.dropna(subset=['artTitle'])\n", "ct = ct.dropna(subset=['artContent'])\n", "# 移除網址格式\n", "ct[\"artContent\"] = ct.artContent.apply(\n", " lambda x: re.sub(\"(http|https)://.*\", \"\", x)\n", ")\n", "ct[\"artTitle\"] = ct[\"artTitle\"].apply(\n", " lambda x: re.sub(\"(http|https)://.*\", \"\", x)\n", ")\n", "# 只留下中文字\n", "ct[\"artContent\"] = ct.artContent.apply(\n", " lambda x: re.sub(\"[^\\u4e00-\\u9fa5]+\", \"\", x)\n", ")\n", "ct[\"artTitle\"] = ct[\"artTitle\"].apply(\n", " lambda x: re.sub(\"[^\\u4e00-\\u9fa5]+\", \"\", x)\n", ")\n", "\n", "# 留下 content\n", "ct[\"content\"] = ct[\"artTitle\"] + ct[\"artContent\"]\n", "ct = ct[[\"content\", \"artUrl\", \"artCatagory\"]] # 文章內容 文章連結\n", "ct.head()\n" ] }, { "cell_type": "code", "execution_count": 102, "id": "755411b3", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "artCatagory\n", "GetMarry 2301\n", "Gossiping 1508\n", "marriage 1347\n", "Name: count, dtype: int64" ] }, "execution_count": 102, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ct['artCatagory'].value_counts()" ] }, { "cell_type": "code", "execution_count": 103, "id": "186f0382", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
contentartUrlartCatagoryembeddings
0推薦彰化客製手工喜餅廠商所在地區彰化是什麼場合用到訂結婚廠商名稱喜餅手工喜餅聯絡資訊官網訂購...https://www.ptt.cc/bbs/GetMarry/M.1653965335.A...GetMarry[0.5550073, -0.20386253, -0.3193656, 0.1848941...
1分享台北荳蔻攝影工作室婚紗照剛開始查婚紗資訊真的好討厭傳統婚紗店的組數限制或是各種無止境加購...https://www.ptt.cc/bbs/GetMarry/M.1653974818.A...GetMarry[0.50719243, -0.0016513392, -0.3523839, 0.2919...
2請益喜餅品項選擇選擇障礙的新娘來求助了區的選擇掙扎到要給品項的目前確定抹茶鹽之花另一個想要檸...https://www.ptt.cc/bbs/GetMarry/M.1654041655.A...GetMarry[0.5617253, -0.14796321, -0.42836052, 0.094054...
\n", "
" ], "text/plain": [ " content \\\n", "0 推薦彰化客製手工喜餅廠商所在地區彰化是什麼場合用到訂結婚廠商名稱喜餅手工喜餅聯絡資訊官網訂購... \n", "1 分享台北荳蔻攝影工作室婚紗照剛開始查婚紗資訊真的好討厭傳統婚紗店的組數限制或是各種無止境加購... \n", "2 請益喜餅品項選擇選擇障礙的新娘來求助了區的選擇掙扎到要給品項的目前確定抹茶鹽之花另一個想要檸... \n", "\n", " artUrl artCatagory \\\n", "0 https://www.ptt.cc/bbs/GetMarry/M.1653965335.A... GetMarry \n", "1 https://www.ptt.cc/bbs/GetMarry/M.1653974818.A... GetMarry \n", "2 https://www.ptt.cc/bbs/GetMarry/M.1654041655.A... GetMarry \n", "\n", " embeddings \n", "0 [0.5550073, -0.20386253, -0.3193656, 0.1848941... \n", "1 [0.50719243, -0.0016513392, -0.3523839, 0.2919... \n", "2 [0.5617253, -0.14796321, -0.42836052, 0.094054... " ] }, "execution_count": 103, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ct[\"embeddings\"] = ct.content.apply(lambda x: bert_ch.encode(x))\n", "ct.head(3)" ] }, { "cell_type": "code", "execution_count": 104, "id": "b2934c33", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " precision recall f1-score support\n", "\n", " GetMarry 1.00 0.99 1.00 2301\n", " Gossiping 0.99 0.97 0.98 1508\n", " marriage 0.97 0.99 0.98 1347\n", "\n", " accuracy 0.99 5156\n", " macro avg 0.99 0.99 0.99 5156\n", "weighted avg 0.99 0.99 0.99 5156\n", "\n" ] } ], "source": [ "X = ct[\"embeddings\"].apply(pd.Series)\n", "y = ct['artCatagory']\n", "\n", "y_pred = clf.predict(X)\n", "print(classification_report(y, y_pred))" ] }, { "cell_type": "code", "execution_count": 109, "id": "72aebfa9", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
artCatagorypred
82GetMarrymarriage
968GetMarryGossiping
975GetMarrymarriage
1103GetMarryGossiping
1108GetMarryGossiping
1168GetMarrymarriage
1396GetMarryGossiping
1789GetMarryGossiping
1948GetMarrymarriage
2046GetMarryGossiping
2109GetMarrymarriage
2176GetMarrymarriage
2381Gossipingmarriage
2437Gossipingmarriage
2454Gossipingmarriage
2456Gossipingmarriage
2465GossipingGetMarry
2466Gossipingmarriage
2474Gossipingmarriage
2545Gossipingmarriage
2607Gossipingmarriage
2646GossipingGetMarry
2718Gossipingmarriage
2838Gossipingmarriage
2860Gossipingmarriage
2896Gossipingmarriage
2946Gossipingmarriage
2992Gossipingmarriage
3221Gossipingmarriage
3226GossipingGetMarry
3317Gossipingmarriage
3383Gossipingmarriage
3408GossipingGetMarry
3425Gossipingmarriage
3428Gossipingmarriage
3501Gossipingmarriage
3521GossipingGetMarry
3566GossipingGetMarry
3615Gossipingmarriage
3617GossipingGetMarry
3630Gossipingmarriage
3657Gossipingmarriage
3664Gossipingmarriage
3669Gossipingmarriage
3671GossipingGetMarry
3714Gossipingmarriage
3757Gossipingmarriage
3775Gossipingmarriage
3801Gossipingmarriage
3805Gossipingmarriage
3855marriageGossiping
4180marriageGossiping
4510marriageGossiping
4532marriageGossiping
4773marriageGossiping
4846marriageGossiping
5226marriageGetMarry
\n", "
" ], "text/plain": [ " artCatagory pred\n", "82 GetMarry marriage\n", "968 GetMarry Gossiping\n", "975 GetMarry marriage\n", "1103 GetMarry Gossiping\n", "1108 GetMarry Gossiping\n", "1168 GetMarry marriage\n", "1396 GetMarry Gossiping\n", "1789 GetMarry Gossiping\n", "1948 GetMarry marriage\n", "2046 GetMarry Gossiping\n", "2109 GetMarry marriage\n", "2176 GetMarry marriage\n", "2381 Gossiping marriage\n", "2437 Gossiping marriage\n", "2454 Gossiping marriage\n", "2456 Gossiping marriage\n", "2465 Gossiping GetMarry\n", "2466 Gossiping marriage\n", "2474 Gossiping marriage\n", "2545 Gossiping marriage\n", "2607 Gossiping marriage\n", "2646 Gossiping GetMarry\n", "2718 Gossiping marriage\n", "2838 Gossiping marriage\n", "2860 Gossiping marriage\n", "2896 Gossiping marriage\n", "2946 Gossiping marriage\n", "2992 Gossiping marriage\n", "3221 Gossiping marriage\n", "3226 Gossiping GetMarry\n", "3317 Gossiping marriage\n", "3383 Gossiping marriage\n", "3408 Gossiping GetMarry\n", "3425 Gossiping marriage\n", "3428 Gossiping marriage\n", "3501 Gossiping marriage\n", "3521 Gossiping GetMarry\n", "3566 Gossiping GetMarry\n", "3615 Gossiping marriage\n", "3617 Gossiping GetMarry\n", "3630 Gossiping marriage\n", "3657 Gossiping marriage\n", "3664 Gossiping marriage\n", "3669 Gossiping marriage\n", "3671 Gossiping GetMarry\n", "3714 Gossiping marriage\n", "3757 Gossiping marriage\n", "3775 Gossiping marriage\n", "3801 Gossiping marriage\n", "3805 Gossiping marriage\n", "3855 marriage Gossiping\n", "4180 marriage Gossiping\n", "4510 marriage Gossiping\n", "4532 marriage Gossiping\n", "4773 marriage Gossiping\n", "4846 marriage Gossiping\n", "5226 marriage GetMarry" ] }, "execution_count": 109, "metadata": {}, "output_type": "execute_result" } ], "source": [ "false_pred = ct.query(\"artCatagory != pred\").loc[:,['artCatagory',\"pred\"]]\n", "false_pred" ] }, { "cell_type": "code", "execution_count": 110, "id": "e406f543", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
artCatagorypred
2381Gossipingmarriage
2437Gossipingmarriage
2454Gossipingmarriage
2456Gossipingmarriage
2465GossipingGetMarry
2466Gossipingmarriage
2474Gossipingmarriage
2545Gossipingmarriage
2607Gossipingmarriage
2646GossipingGetMarry
2718Gossipingmarriage
2838Gossipingmarriage
2860Gossipingmarriage
2896Gossipingmarriage
2946Gossipingmarriage
2992Gossipingmarriage
3221Gossipingmarriage
3226GossipingGetMarry
3317Gossipingmarriage
3383Gossipingmarriage
\n", "
" ], "text/plain": [ " artCatagory pred\n", "2381 Gossiping marriage\n", "2437 Gossiping marriage\n", "2454 Gossiping marriage\n", "2456 Gossiping marriage\n", "2465 Gossiping GetMarry\n", "2466 Gossiping marriage\n", "2474 Gossiping marriage\n", "2545 Gossiping marriage\n", "2607 Gossiping marriage\n", "2646 Gossiping GetMarry\n", "2718 Gossiping marriage\n", "2838 Gossiping marriage\n", "2860 Gossiping marriage\n", "2896 Gossiping marriage\n", "2946 Gossiping marriage\n", "2992 Gossiping marriage\n", "3221 Gossiping marriage\n", "3226 Gossiping GetMarry\n", "3317 Gossiping marriage\n", "3383 Gossiping marriage" ] }, "execution_count": 110, "metadata": {}, "output_type": "execute_result" } ], "source": [ "false_pred.loc[false_pred['artCatagory']=='Gossiping', :].head(20)" ] }, { "cell_type": "code", "execution_count": 115, "id": "cdae96d4", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
artCatagorypredcontentembeddings
2381Gossipingmarriage問卦少子化女生的問題佔大部分吧台女白天要上班下班要養家假日要讀書沒辦法到外面聯誼之前我在臉書...[0.8725247, -0.08411743, -0.203732, 0.27536434...
2437Gossipingmarriage新聞工程師尪被孕妻抓包上按摩店性交易簽工程師尪被孕妻抓包上按摩店性交易簽完協議再犯下場曝光年...[0.46804097, -0.016441723, -0.2565862, 0.06258...
2454Gossipingmarriage問卦你各位買得起房就真的會結婚生小孩結婚可能但生小孩很難買不起房然後呢別忘了還有房貸再來是小...[0.65234625, -0.09828554, -0.12775144, 0.26947...
2456Gossipingmarriage問卦台女找尋結婚對象有人留言說不生幹嘛結婚我只是能說太天真了婚姻是個強大道德武器之後至於要不...[0.79514945, -0.067468375, -0.14028932, 0.4278...
2465GossipingGetMarry問卦台女找尋結婚對象小妹我的親辜徵友條件跟低卡上這位列的有相似但小妹的親辜身高體重而且比較年...[0.928948, -0.1616967, 0.04163605, 0.56533563,...
2466Gossipingmarriage問卦家長是巨嬰嗎一般來說親子關係是要培養的不管是爸爸還是媽媽不是只負責生不負責養那以前都會說...[0.8342646, -0.11413084, -0.100559324, 0.21266...
2474Gossipingmarriage新聞色人妻新婚不久就出軌誘惑男同事小色人妻新婚不久就出軌誘惑男同事小朋友才戴套把你榨乾記者柯...[0.58158386, 0.022150613, -0.3089661, 0.124260...
2545Gossipingmarriage問卦老婆只想生一個建議你先跟你老婆借一下手機然後把你老婆的手機格式化再跟她講解備份的重要性有...[0.5313501, -0.26143715, -0.0760047, 0.5591018...
2607Gossipingmarriage問卦男人不婚不生把錢花在出國也爽以前常有人說台女沒啥在存錢總是把錢花在出國玩樂享受美景美食上...[0.5529267, 0.08784643, -0.022090623, 0.515583...
2646GossipingGetMarry問卦有萬為什麼不娶台灣要娶越南老婆對聽說去年有位造型師娶越南老婆但是價格驚人全部流程加上婚禮...[0.5280451, -0.01208675, 0.1979833, 0.43109772...
\n", "
" ], "text/plain": [ " artCatagory pred content \\\n", "2381 Gossiping marriage 問卦少子化女生的問題佔大部分吧台女白天要上班下班要養家假日要讀書沒辦法到外面聯誼之前我在臉書... \n", "2437 Gossiping marriage 新聞工程師尪被孕妻抓包上按摩店性交易簽工程師尪被孕妻抓包上按摩店性交易簽完協議再犯下場曝光年... \n", "2454 Gossiping marriage 問卦你各位買得起房就真的會結婚生小孩結婚可能但生小孩很難買不起房然後呢別忘了還有房貸再來是小... \n", "2456 Gossiping marriage 問卦台女找尋結婚對象有人留言說不生幹嘛結婚我只是能說太天真了婚姻是個強大道德武器之後至於要不... \n", "2465 Gossiping GetMarry 問卦台女找尋結婚對象小妹我的親辜徵友條件跟低卡上這位列的有相似但小妹的親辜身高體重而且比較年... \n", "2466 Gossiping marriage 問卦家長是巨嬰嗎一般來說親子關係是要培養的不管是爸爸還是媽媽不是只負責生不負責養那以前都會說... \n", "2474 Gossiping marriage 新聞色人妻新婚不久就出軌誘惑男同事小色人妻新婚不久就出軌誘惑男同事小朋友才戴套把你榨乾記者柯... \n", "2545 Gossiping marriage 問卦老婆只想生一個建議你先跟你老婆借一下手機然後把你老婆的手機格式化再跟她講解備份的重要性有... \n", "2607 Gossiping marriage 問卦男人不婚不生把錢花在出國也爽以前常有人說台女沒啥在存錢總是把錢花在出國玩樂享受美景美食上... \n", "2646 Gossiping GetMarry 問卦有萬為什麼不娶台灣要娶越南老婆對聽說去年有位造型師娶越南老婆但是價格驚人全部流程加上婚禮... \n", "\n", " embeddings \n", "2381 [0.8725247, -0.08411743, -0.203732, 0.27536434... \n", "2437 [0.46804097, -0.016441723, -0.2565862, 0.06258... \n", "2454 [0.65234625, -0.09828554, -0.12775144, 0.26947... \n", "2456 [0.79514945, -0.067468375, -0.14028932, 0.4278... \n", "2465 [0.928948, -0.1616967, 0.04163605, 0.56533563,... \n", "2466 [0.8342646, -0.11413084, -0.100559324, 0.21266... \n", "2474 [0.58158386, 0.022150613, -0.3089661, 0.124260... \n", "2545 [0.5313501, -0.26143715, -0.0760047, 0.5591018... \n", "2607 [0.5529267, 0.08784643, -0.022090623, 0.515583... \n", "2646 [0.5280451, -0.01208675, 0.1979833, 0.43109772... " ] }, "execution_count": 115, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ct_wrong = ct[(ct[\"artCatagory\"] == \"Gossiping\") & (ct[\"pred\"] != \"Gossiping\")]\n", "ct_wrong[[\"artCatagory\", \"pred\", \"content\",\"embeddings\"]].head(10)\n" ] }, { "cell_type": "markdown", "id": "5aae1db1", "metadata": {}, "source": [ "## 第三次讀書會 BERT (Encoder-only-model)\n", "---" ] }, { "cell_type": "markdown", "id": "c7c09a4c", "metadata": {}, "source": [ "**前處理常用套件**\n" ] }, { "cell_type": "code", "execution_count": null, "id": "ddc61652", "metadata": {}, "outputs": [], "source": [ "!pip install jieba" ] }, { "cell_type": "code", "execution_count": null, "id": "9e0abbd2", "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "import re\n", "import numpy as np\n", "from collections import defaultdict\n", "import multiprocessing\n", "import jieba\n", "import matplotlib.pyplot as plt\n", "from matplotlib.font_manager import fontManager\n", "\n", "# 設定字體\n", "fontManager.addfont('./TaipeiSansTCBeta-Regular.ttf')\n", "plt.rcParams['font.sans-serif'] = ['Taipei Sans TC Beta']\n", "plt.rcParams['font.size'] = '16'" ] }, { "cell_type": "markdown", "id": "e2ab3234", "metadata": {}, "source": [ "**Transformers 和 Sentence-transformers （使用 huggingface 模型）**" ] }, { "cell_type": "code", "execution_count": null, "id": "8b6b7b4d", "metadata": {}, "outputs": [], "source": [ "!pip install sentence_transformers\n", "!pip install ckip_transformers" ] }, { "cell_type": "code", "execution_count": null, "id": "5f1e1edd", "metadata": {}, "outputs": [], "source": [ "from transformers import BertTokenizerFast, AutoTokenizer, AutoModelForTokenClassification, AutoModelForSequenceClassification, pipeline\n", "from sentence_transformers import SentenceTransformer\n", "from ckip_transformers.nlp import CkipWordSegmenter, CkipPosTagger, CkipNerChunker" ] }, { "cell_type": "markdown", "id": "fd53b4c2", "metadata": {}, "source": [ "**BERTopic套件**" ] }, { "cell_type": "code", "execution_count": null, "id": "c14979a8", "metadata": {}, "outputs": [], "source": [ "!pip install bertopic" ] }, { "cell_type": "code", "execution_count": null, "id": "59cd4033", "metadata": {}, "outputs": [], "source": [ "from bertopic import BERTopic\n", "from bertopic.vectorizers import ClassTfidfTransformer\n", "from hdbscan import HDBSCAN\n", "from sklearn.feature_extraction.text import CountVectorizer\n", "from sklearn.cluster import KMeans" ] }, { "cell_type": "markdown", "id": "9bd0c6ae", "metadata": {}, "source": [ "### 2. 資料前處理" ] }, { "cell_type": "markdown", "id": "e28698a4", "metadata": {}, "source": [ "中文資料集：載入離婚資料集" ] }, { "cell_type": "code", "execution_count": null, "id": "1b699d88", "metadata": {}, "outputs": [], "source": [ "# 讀入中文示範資料集\n", "# origin_data = pd.read_csv('./raw_data/zh_buffet_20_22.csv')\n", "origin_data = pd.read_csv('./raw_data/text_marriage.csv')" ] }, { "cell_type": "code", "execution_count": null, "id": "a261617a", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
system_idartUrlartTitleartDateartContentsentence
01https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...常常看大家說
11https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...偷看手機是不對的
21https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...但如果已經結婚了
31https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...太太想看你手機
41https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...真的可以拒絕嗎
51https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...感覺你拒絕
61https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...就是心裡有鬼
71https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...只是讓太太猜忌
81https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...自己日子難過
91https://www.ptt.cc/bbs/marriage/M.1610159827.A...[求助]真的可以不給看手機嗎？2021-01-09 10:37:05常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\...長期以來
\n", "
" ], "text/plain": [ " system_id artUrl \\\n", "0 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "1 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "2 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "3 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "4 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "5 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "6 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "7 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "8 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "9 1 https://www.ptt.cc/bbs/marriage/M.1610159827.A... \n", "\n", " artTitle artDate \\\n", "0 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "1 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "2 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "3 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "4 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "5 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "6 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "7 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "8 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "9 [求助]真的可以不給看手機嗎？ 2021-01-09 10:37:05 \n", "\n", " artContent sentence \n", "0 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 常常看大家說 \n", "1 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 偷看手機是不對的 \n", "2 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 但如果已經結婚了 \n", "3 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 太太想看你手機 \n", "4 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 真的可以拒絕嗎 \n", "5 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 感覺你拒絕 \n", "6 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 就是心裡有鬼 \n", "7 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 只是讓太太猜忌 \n", "8 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 自己日子難過 \n", "9 常常看大家說，偷看手機是不對的，\\n但如果已經結婚了，太太想看你手機，\\n真的可以拒絕嗎？\\... 長期以來 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 去除一些不需要的欄位\n", "metaData = origin_data.drop(['artPoster', 'artCatagory', 'artComment', 'e_ip', 'insertedDate', 'dataSource'], axis=1)\n", "\n", "# 只留下中文字\n", "metaData['sentence'] = metaData['artContent'].str.replace(r'\\n\\n','。', regex=True)\n", "metaData['sentence'] = metaData['sentence'].str.replace(r'\\n','', regex=True)\n", "\n", "metaData['sentence'] = metaData['sentence'].str.split(\"[,，。！!？?]{1,}\")\n", "metaData = metaData.explode('sentence').reset_index(drop=True)\n", "\n", "metaData['sentence'] = metaData['sentence'].apply(lambda x: re.sub('[^\\u4e00-\\u9fff]+', '',x))\n", "\n", "metaData.head(10)" ] }, { "cell_type": "markdown", "id": "4fb7a426", "metadata": {}, "source": [ "## 3. Token classification" ] }, { "cell_type": "markdown", "id": "7a49f49f", "metadata": {}, "source": [ "### NER\n", "使用 Huggingface 上面已經針對 NER 任務 finetune 好的 BERT 模型來實作
\n", "Huggingface 的模型列表：https://huggingface.co/models?sort=trending" ] }, { "cell_type": "markdown", "id": "531530fc", "metadata": {}, "source": [ "#### 3.1 中文 NER：
\n", "- 使用套件：transformers
\n", "- 使用的 NER 模型：https://huggingface.co/ckiplab/bert-base-chinese-ner" ] }, { "cell_type": "code", "execution_count": null, "id": "6fe27d64", "metadata": {}, "outputs": [], "source": [ "# 載入中文NER模型\n", "model_name = 'ckiplab/bert-base-chinese-ner'\n", "tokenizer = BertTokenizerFast.from_pretrained(model_name)\n", "model = AutoModelForTokenClassification.from_pretrained(model_name)" ] }, { "cell_type": "markdown", "id": "8cadce14", "metadata": {}, "source": [ "也可以使用 CKIP 開發的 NLP 套件：ckip_transformers
\n", "- 使用的 WS 模型：https://huggingface.co/ckiplab/bert-base-chinese-ws
\n", "- 使用的 POS 模型：https://huggingface.co/ckiplab/bert-base-chinese-pos
\n", "- 使用的 NER 模型：https://huggingface.co/ckiplab/bert-base-chinese-ner" ] }, { "cell_type": "code", "execution_count": null, "id": "91f1cd65", "metadata": {}, "outputs": [], "source": [ "# 初始化 ckip 工具 device=0 使用GPU ｜ device=-1 使用CPU（速度會很慢）\n", "# Mac使用者可以設定 device=torch.device(\"mps\") 使用GPU\n", "ws_driver = CkipWordSegmenter(model_name=\"ckiplab/bert-base-chinese-ws\", device=0) # Word Segmenter斷詞\n", "pos_driver = CkipPosTagger(model_name=\"ckiplab/bert-base-chinese-pos\", device=0) # POS tagger 詞性標記\n", "ner_driver = CkipNerChunker(model_name=\"ckiplab/bert-base-chinese-ner\", device=0) # NER識別\n" ] }, { "cell_type": "markdown", "id": "65fd99d1", "metadata": {}, "source": [ "**將CKIP套用到我們先前處理好的資料集**" ] }, { "cell_type": "code", "execution_count": null, "id": "0ad6ce38", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Tokenization: 100%|██████████| 50/50 [00:00\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
sentencepacked_sentenceentities
0常常看大家說常常(D)　看(VC)　大家(Nh)　說(VE)[]
1偷看手機是不對的偷看(VC)　手機(Na)　是(SHI)　不(D)　對(VH)　的(DE)[]
2但如果已經結婚了但(Cbb)　如果(Cbb)　已經(D)　結婚(VA)　了(Di)[]
3太太想看你手機太太(Na)　想(VE)　看(VC)　你(Nh)　手機(Na)[]
4真的可以拒絕嗎真的(D)　可以(D)　拒絕(VF)　嗎(T)[]
5感覺你拒絕感覺(VK)　你(Nh)　拒絕(VF)[]
6就是心裡有鬼就(D)　是(SHI)　心(Na)　裡(Ng)　有(V_2)　鬼(Na)[]
7只是讓太太猜忌只是(D)　讓(VL)　太太(Na)　猜忌(VJ)[]
8自己日子難過自己(Nh)　日子(Na)　難過(VK)[]
9長期以來長期(Na)　以來(Ng)[]
\n", "" ], "text/plain": [ " sentence packed_sentence entities\n", "0 常常看大家說 常常(D)　看(VC)　大家(Nh)　說(VE) []\n", "1 偷看手機是不對的 偷看(VC)　手機(Na)　是(SHI)　不(D)　對(VH)　的(DE) []\n", "2 但如果已經結婚了 但(Cbb)　如果(Cbb)　已經(D)　結婚(VA)　了(Di) []\n", "3 太太想看你手機 太太(Na)　想(VE)　看(VC)　你(Nh)　手機(Na) []\n", "4 真的可以拒絕嗎 真的(D)　可以(D)　拒絕(VF)　嗎(T) []\n", "5 感覺你拒絕 感覺(VK)　你(Nh)　拒絕(VF) []\n", "6 就是心裡有鬼 就(D)　是(SHI)　心(Na)　裡(Ng)　有(V_2)　鬼(Na) []\n", "7 只是讓太太猜忌 只是(D)　讓(VL)　太太(Na)　猜忌(VJ) []\n", "8 自己日子難過 自己(Nh)　日子(Na)　難過(VK) []\n", "9 長期以來 長期(Na)　以來(Ng) []" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 以前50筆資料作為範例\n", "text = metaData['sentence'].tolist()\n", "text = text[:50]\n", "\n", "# 執行處理 \n", "ws = ws_driver(text) # 斷詞\n", "pos = pos_driver(ws) # POS\n", "ner = ner_driver(text) # NER\n", "\n", "# 將斷詞以及 pos 結果合在一起顯示\n", "def pack_ws_pos_sentece(sentence_ws, sentence_pos):\n", " assert len(sentence_ws) == len(sentence_pos) # 確認斷詞和POS的長度相同\n", " res = []\n", " for word_ws, word_pos in zip(sentence_ws, sentence_pos):\n", " res.append(f\"{word_ws}({word_pos})\") # 合併在一起\n", " return \"\\u3000\".join(res) \n", "\n", "sentences, packed_sentences, entities = [], [], []\n", "\n", "# 儲存結果\n", "for sentence, sentence_ws, sentence_pos, sentence_ner in zip(text, ws, pos, ner):\n", " sentences.append(sentence)\n", " packed_sentences.append(pack_ws_pos_sentece(sentence_ws, sentence_pos))\n", " entities.append([str(entity) for entity in sentence_ner])\n", "\n", "# 將結果存在一個 dataframe 中\n", "ner_results = pd.DataFrame({\n", " 'sentence': sentences,\n", " 'packed_sentence': packed_sentences,\n", " 'entities': entities\n", "})\n", "\n", "ner_results.head(10)\n" ] }, { "cell_type": "markdown", "id": "b6e0f4af", "metadata": {}, "source": [ "用CKIP Transformers提供的模型套件進行斷詞、詞性標註和實體辨識的結果" ] }, { "cell_type": "markdown", "id": "7b281430", "metadata": {}, "source": [ "## 4. Sequence classification" ] }, { "cell_type": "markdown", "id": "a7fb6e18", "metadata": {}, "source": [ "### 4.1 Sentiment Classification\n", "使用 Huggingface 上面已經針對 Sentiment classification 任務 finetune 的 BERT 模型來實作
\n", "使用的模型：https://huggingface.co/techthiyanes/chinese_sentiment

\n", "情緒(start 1到star 5)：
\n", "1. Semi-negation
\n", "2. Negation
\n", "3. Neutral
\n", "4. Semi-positive
\n", "5. Positive" ] }, { "cell_type": "code", "execution_count": null, "id": "0d166115", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Device set to use cuda:0\n" ] } ], "source": [ "# 載入已經被 fine-tune 過的 BERT 模型\n", "model_name = \"techthiyanes/chinese_sentiment\" # 你可以將這裡換成你想要使用的模型\n", "# model = pipeline('sentiment-analysis', model=model_name)\n", "model = pipeline('sentiment-analysis', model=model_name, device=0)\n" ] }, { "cell_type": "code", "execution_count": null, "id": "7fb5e035", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
sentencelabelscore
0常常看大家說star 40.509952
1偷看手機是不對的star 20.314251
2但如果已經結婚了star 30.319404
3太太想看你手機star 20.288831
4真的可以拒絕嗎star 10.465014
5感覺你拒絕star 50.336695
6就是心裡有鬼star 20.328898
7只是讓太太猜忌star 20.311495
8自己日子難過star 10.475942
9長期以來star 40.445570
\n", "
" ], "text/plain": [ " sentence label score\n", "0 常常看大家說 star 4 0.509952\n", "1 偷看手機是不對的 star 2 0.314251\n", "2 但如果已經結婚了 star 3 0.319404\n", "3 太太想看你手機 star 2 0.288831\n", "4 真的可以拒絕嗎 star 1 0.465014\n", "5 感覺你拒絕 star 5 0.336695\n", "6 就是心裡有鬼 star 2 0.328898\n", "7 只是讓太太猜忌 star 2 0.311495\n", "8 自己日子難過 star 1 0.475942\n", "9 長期以來 star 4 0.445570" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# # 建立一個新的 dataframe 來儲存結果\n", "# results_df = pd.DataFrame(columns=['sentence', 'label', 'score'])\n", "# results_df['sentence'] = metaData['sentence']\n", "\n", "# # 定義一個函數來進行情緒分析\n", "# # def analyze_sentiment(sentence):\n", "# # result = model([sentence])\n", "# # return pd.Series([result[0]['label'], result[0]['score']])\n", "# def analyze_sentiment(sentence):\n", "# result = model(sentence, truncation=True, max_length=512, batch_size=8) # 用 tokenizer 算 token 數、切割長度，就不用再用 x[:500] 去砍字元\n", "# return pd.Series([result[0]['label'], result[0]['score']])\n", "\n", "\n", "# # 使用 apply 函數來進行情緒分析\n", "# # metaData['sentence'] = metaData['sentence'].apply(lambda x: x[:500])\n", "# results_df[['label', 'score']] = metaData['sentence'].apply(analyze_sentiment)\n", "\n", "# # 輸出結果\n", "# results_df.head(10)\n", "\n", "# 取出所有句子\n", "sentences = metaData['sentence'].tolist()\n", "\n", "# 批次跑 pipeline（一次丟入全部，內部分批跑 batch_size）\n", "results = model(sentences, truncation=True, max_length=512, batch_size=8)\n", "\n", "# 組成結果 DataFrame\n", "results_df = pd.DataFrame({\n", " 'sentence': sentences,\n", " 'label': [res['label'] for res in results],\n", " 'score': [res['score'] for res in results]\n", "})\n", "\n", "# 輸出結果\n", "results_df.head(10)\n", "\n" ] }, { "cell_type": "markdown", "id": "8f5f2732", "metadata": {}, "source": [ "如\"常常看大家說\"被分在star 4，是較為正向的，而\"自己日子難過\"是被分在star 1，是最負向" ] }, { "cell_type": "markdown", "id": "b14a1bdd", "metadata": {}, "source": [ "## 5. Text Clustering" ] }, { "cell_type": "markdown", "id": "bb120602", "metadata": {}, "source": [ "#### 套用於中文
\n", "為了套用到中文文章，各元件必須修改為支援中文的方法，主要針對 embedding model 以及 tokenizer" ] }, { "cell_type": "code", "execution_count": null, "id": "a378ecfe", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "103443\n" ] } ], "source": [ "print(len(metaData))" ] }, { "cell_type": "code", "execution_count": null, "id": "60d99703", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "No sentence-transformers model found with name google-bert/bert-base-chinese. Creating a new one with mean pooling.\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Encoding batch 1 / 21\n" ] }, { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "f33dc7ca51604dbaa535a643089a49ac", "version_major": 2, "version_minor": 0 }, "text/plain": [ "Batches: 0%| | 0/157 [00:00\n", "(103443, 768)\n" ] } ], "source": [ "print(type(all_embeddings))\n", "print(all_embeddings.shape)" ] }, { "cell_type": "code", "execution_count": null, "id": "bff28fe9", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "2025-05-10 03:34:32,454 - BERTopic - Dimensionality - Fitting the dimensionality reduction algorithm\n", "2025-05-10 03:35:01,087 - BERTopic - Dimensionality - Completed ✓\n", "2025-05-10 03:35:01,088 - BERTopic - Cluster - Start clustering the reduced embeddings\n", "2025-05-10 03:35:05,307 - BERTopic - Cluster - Completed ✓\n", "2025-05-10 03:35:05,315 - BERTopic - Representation - Fine-tuning topics using representation models.\n", "2025-05-10 03:35:09,015 - BERTopic - Representation - Completed ✓\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ " Topic Count Name \\\n", "0 -1 77800 -1_ _說_小孩_妳 \n", "1 0 7361 0_ _老婆_妳_婚姻 \n", "2 1 6016 1_ _真的_說_情況 \n", "3 2 4244 2_房子_ _房貸_萬 \n", "4 3 3396 3_ _小孩_睡_下班 \n", "5 4 2522 4_ _文章_版友_建議 \n", "6 5 2104 5_離婚_ _結婚_外遇 \n", "\n", " Representation \\\n", "0 [ , 說, 小孩, 妳, 老婆, 想, 老公, 真的, 做, 離婚, 孩子, 工作, 太太... \n", "1 [ , 老婆, 妳, 婚姻, 老公, 做, 說, 離婚, 改變, 想, 人生, 小孩, 真的... \n", "2 [ , 真的, 說, 情況, 想, 太, 答案, 妳, 不想, 方法, 痛苦, 做, 喔, ... \n", "3 [房子, , 房貸, 萬, 錢, 薪水, 財產, 買, 工作, 貸款, 費用, 買房, 收... \n", "4 [ , 小孩, 睡, 下班, 回家, 吃, 晚上, 洗, 時間, 睡覺, 假日, 煮, 上班... \n", "5 [ , 文章, 版友, 建議, 謝謝, 發文, 分享, 推文, 參考, 想, 文, 回文, ... \n", "6 [離婚, , 結婚, 外遇, 交往, 想, 提, 分手, 建議, 真的, 說, 談, 不想... \n", "\n", " Representative_Docs \n", "0 [但有小孩的話, 阿不然你說說他有哪些好, 而不是忽然你父母說要帶小孩] \n", "1 [最後我想說的是我懂妳那種在一個地方孤立無援的感覺如果跟先生一直找不到共識那就自己先做出改變... \n", "2 [我說, 像大說的, 那真的先不要] \n", "3 [而房子就歸我了, 房子和, 沒人知道房子是我的] \n", "4 [他想好好休息我陪睡的時候我也不會拒絕我都是哄完小孩睡覺才會去睡要小時顧才不廢, 媽媽自己做... \n", "5 [你的文章只看到, 看了你文章的推文, 感謝大家的回覆與建議沒想到會這麼多人回覆早知道很多年... \n", "6 [也是直接離婚就好, 趕快離婚, 也差不多可以離婚了] \n" ] } ], "source": [ "\n", "\n", "# ---------- 定義 jieba 分詞函數 ----------\n", "def tokenize_zh(text):\n", " words = jieba.lcut(text)\n", " return words\n", "\n", "# ---------- 設定 CountVectorizer（含 jieba 分詞 + 停用詞） ----------\n", "jieba_vectorizer = CountVectorizer(\n", " tokenizer=tokenize_zh,\n", " stop_words=stopwords,\n", " analyzer='word',\n", " token_pattern=u\"(?u)\\\\b\\\\w+\\\\b\"\n", ")\n", "\n", "# ---------- 設定 HDBSCAN 參數（可調整群組靈敏度） ----------\n", "hdbscan_model = HDBSCAN(min_cluster_size=2000, min_samples=10)\n", "\n", "# ---------- 建立 BERTopic 模型 ----------\n", "zh_topic_model = BERTopic(\n", " embedding_model=bert_sentence_model,\n", " vectorizer_model=jieba_vectorizer,\n", " hdbscan_model=hdbscan_model,\n", " verbose=True,\n", " top_n_words=30\n", ")\n", "\n", "# ---------- 跑主題模型 ----------\n", "topics, probs = zh_topic_model.fit_transform(docs_zh, all_embeddings)\n", "\n", "# ---------- 查看主題資訊 ----------\n", "topic_info = zh_topic_model.get_topic_info()\n", "print(topic_info)\n" ] }, { "cell_type": "markdown", "id": "3551b3f0", "metadata": {}, "source": [ "總共分出七個主題，其中第一個主題無法分類的詞彙，第二個則與婚姻中的互動關係較有關，包含老公、老婆、小孩、婚姻、離婚等詞彙，第三個則與較抽象的情感表達詞彙有關，如真的、不想、痛苦等，第四個則與房子、經濟、財務相關，主要是針對經濟壓力、買房貸款、家庭財務分配的討論，第五個則與孩子及家庭生活有關，有較多關於照顧孩子、作息安排與家庭生活的描述，第六個則是與社群平台的互動行為較相關，就與我們的主題較無關聯，但有被單獨分成一類，第七個則與感情、婚姻破裂較相關，有外遇、分手、離婚等詞，多是關於分手、離婚、外遇的感情問題。
\n", "所以是有成功辨識出多個的明確主題，例如「婚姻問題」、「房貸壓力」、「家庭生活」。" ] }, { "cell_type": "code", "execution_count": null, "id": "5cbbd497", "metadata": {}, "outputs": [ { "data": { "application/vnd.plotly.v1+json": { "config": { "plotlyServerURL": "https://plot.ly" }, "data": [ { "customdata": [ [ 0, " | 老婆 | 妳 | 婚姻 | 老公", 7361 ], [ 1, " | 真的 | 說 | 情況 | 想", 6016 ], [ 2, "房子 | | 房貸 | 萬 | 錢", 4244 ], [ 3, " | 小孩 | 睡 | 下班 | 回家", 3396 ], [ 4, " | 文章 | 版友 | 建議 | 謝謝", 2522 ], [ 5, "離婚 | | 結婚 | 外遇 | 交往", 2104 ] ], "hovertemplate": "Topic %{customdata[0]}
%{customdata[1]}
Size: %{customdata[2]}", "legendgroup": "", "marker": { "color": "#B0BEC5", "line": { "color": "DarkSlateGrey", "width": 2 }, "size": { "bdata": "wRyAF5QQRA3aCTgI", "dtype": "i2" }, "sizemode": "area", "sizeref": 4.600625, "symbol": "circle" }, "mode": "markers", "name": "", "orientation": "v", "showlegend": false, "type": "scatter", "x": { "bdata": "m7ihQa5TpEHz0ptBxOCqQcOUo0FiCJ9B", "dtype": "f4" }, "xaxis": "x", "y": { "bdata": "SkVEwTtPNcGM30rBo/lawfSWTsEF4zfB", "dtype": "f4" }, "yaxis": "y" } ], "layout": { "annotations": [ { "showarrow": false, "text": "D1", "x": 16.55630216598511, "y": -12.685452651977538, "yshift": 10 }, { "showarrow": false, "text": "D2", "x": 20.560006666183472, "xshift": 10, "y": -9.632066869735718 } ], "height": 650, "hoverlabel": { "bgcolor": "white", "font": { "family": "Rockwell", "size": 16 } }, "legend": { "itemsizing": "constant", "tracegroupgap": 0 }, "margin": { "t": 60 }, "shapes": [ { "line": { "color": "#CFD8DC", "width": 2 }, "type": "line", "x0": 20.560006666183472, "x1": 20.560006666183472, "y0": -15.73883843421936, "y1": -9.632066869735718 }, { "line": { "color": "#9E9E9E", "width": 2 }, "type": "line", "x0": 16.55630216598511, "x1": 24.563711166381836, "y0": -12.685452651977538, "y1": -12.685452651977538 } ], "sliders": [ { "active": 0, "pad": { "t": 50 }, "steps": [ { "args": [ { "marker.color": [ [ "red", "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5" ] ] } ], "label": "Topic 0", "method": "update" }, { "args": [ { "marker.color": [ [ "#B0BEC5", "red", "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5" ] ] } ], "label": "Topic 1", "method": "update" }, { "args": [ { "marker.color": [ [ "#B0BEC5", "#B0BEC5", "red", "#B0BEC5", "#B0BEC5", "#B0BEC5" ] ] } ], "label": "Topic 2", "method": "update" }, { "args": [ { "marker.color": [ [ "#B0BEC5", "#B0BEC5", "#B0BEC5", "red", "#B0BEC5", "#B0BEC5" ] ] } ], "label": "Topic 3", "method": "update" }, { "args": [ { "marker.color": [ [ "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5", "red", "#B0BEC5" ] ] } ], "label": "Topic 4", "method": "update" }, { "args": [ { "marker.color": [ [ "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5", "#B0BEC5", "red" ] ] } ], "label": "Topic 5", "method": "update" } ] } ], "template": { "data": { "bar": [ { "error_x": { "color": "rgb(36,36,36)" }, "error_y": { "color": "rgb(36,36,36)" }, "marker": { "line": { "color": "white", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "bar" } ], "barpolar": [ { "marker": { "line": { "color": "white", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "barpolar" } ], "carpet": [ { "aaxis": { "endlinecolor": "rgb(36,36,36)", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "rgb(36,36,36)" }, "baxis": { "endlinecolor": "rgb(36,36,36)", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "rgb(36,36,36)" }, "type": "carpet" } ], "choropleth": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "choropleth" } ], "contour": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "contour" } ], "contourcarpet": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "contourcarpet" } ], "heatmap": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "heatmap" } ], "histogram": [ { "marker": { "line": { "color": "white", "width": 0.6 } }, "type": "histogram" } ], "histogram2d": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "histogram2d" } ], "histogram2dcontour": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "histogram2dcontour" } ], "mesh3d": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "mesh3d" } ], "parcoords": [ { "line": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "parcoords" } ], "pie": [ { "automargin": true, "type": "pie" } ], "scatter": [ { "fillpattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 }, "type": "scatter" } ], "scatter3d": [ { "line": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatter3d" } ], "scattercarpet": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattercarpet" } ], "scattergeo": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattergeo" } ], "scattergl": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattergl" } ], "scattermap": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattermap" } ], "scattermapbox": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattermapbox" } ], "scatterpolar": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterpolar" } ], "scatterpolargl": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterpolargl" } ], "scatterternary": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterternary" } ], "surface": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "surface" } ], "table": [ { "cells": { "fill": { "color": "rgb(237,237,237)" }, "line": { "color": "white" } }, "header": { "fill": { "color": "rgb(217,217,217)" }, "line": { "color": "white" } }, "type": "table" } ] }, "layout": { "annotationdefaults": { "arrowhead": 0, "arrowwidth": 1 }, "autotypenumbers": "strict", "coloraxis": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "colorscale": { "diverging": [ [ 0, "rgb(103,0,31)" ], [ 0.1, "rgb(178,24,43)" ], [ 0.2, "rgb(214,96,77)" ], [ 0.3, "rgb(244,165,130)" ], [ 0.4, "rgb(253,219,199)" ], [ 0.5, "rgb(247,247,247)" ], [ 0.6, "rgb(209,229,240)" ], [ 0.7, "rgb(146,197,222)" ], [ 0.8, "rgb(67,147,195)" ], [ 0.9, "rgb(33,102,172)" ], [ 1, "rgb(5,48,97)" ] ], "sequential": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "sequentialminus": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ] }, "colorway": [ "#1F77B4", "#FF7F0E", "#2CA02C", "#D62728", "#9467BD", "#8C564B", "#E377C2", "#7F7F7F", "#BCBD22", "#17BECF" ], "font": { "color": "rgb(36,36,36)" }, "geo": { "bgcolor": "white", "lakecolor": "white", "landcolor": "white", "showlakes": true, "showland": true, "subunitcolor": "white" }, "hoverlabel": { "align": "left" }, "hovermode": "closest", "mapbox": { "style": "light" }, "paper_bgcolor": "white", "plot_bgcolor": "white", "polar": { "angularaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "bgcolor": "white", "radialaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" } }, "scene": { "xaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "yaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "zaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" } }, "shapedefaults": { "fillcolor": "black", "line": { "width": 0 }, "opacity": 0.3 }, "ternary": { "aaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "baxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "bgcolor": "white", "caxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" } }, "title": { "x": 0.05 }, "xaxis": { "automargin": true, "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside", "title": { "standoff": 15 }, "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "yaxis": { "automargin": true, "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside", "title": { "standoff": 15 }, "zeroline": false, "zerolinecolor": "rgb(36,36,36)" } } }, "title": { "font": { "color": "Black", "size": 22 }, "text": "Intertopic Distance Map", "x": 0.5, "xanchor": "center", "y": 0.95, "yanchor": "top" }, "width": 650, "xaxis": { "anchor": "y", "domain": [ 0, 1 ], "range": [ 16.55630216598511, 24.563711166381836 ], "title": { "text": "" }, "visible": false }, "yaxis": { "anchor": "x", "domain": [ 0, 1 ], "range": [ -15.73883843421936, -9.632066869735718 ], "title": { "text": "" }, "visible": false } } } }, "metadata": {}, "output_type": "display_data" } ], "source": [ "zh_topic_model.visualize_topics()" ] }, { "cell_type": "markdown", "id": "81e7bbe0", "metadata": {}, "source": [ "我們將第一個主題去掉，而剩下的主題，從這個分布圖我們可以以看出情緒表達詞彙與婚姻關係破裂的詞會較相關，而與孩子、家庭生活相關的這個主題就與其他主體相關性較低，語意較少與其他主題重疊。" ] }, { "cell_type": "code", "execution_count": null, "id": "3c888d9d", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 104/104 [00:15<00:00, 6.68it/s]\n" ] } ], "source": [ "# 估算每個文件對BERTopic每個主題的機率分布\n", "topic_distr, _ = zh_topic_model.approximate_distribution(docs_zh)" ] }, { "cell_type": "code", "execution_count": null, "id": "c9ad4f31", "metadata": {}, "outputs": [ { "data": { "application/vnd.plotly.v1+json": { "config": { "plotlyServerURL": "https://plot.ly" }, "data": [ { "marker": { "color": "#C8D2D7", "line": { "color": "#6E8484", "width": 1 } }, "orientation": "h", "type": "bar", "x": [ 0.2030228180373761, 0.2904274787816493, 0.13669468061291898, 0.09567379287617377, 0.08618057255827931, 0.18800065713360267 ], "y": [ "Topic 0: _老婆_妳_婚姻_老公", "Topic 1: _真的_說_情況_想", "Topic 2: 房子_ _房貸_萬_錢", "Topic 3: _小孩_睡_下班_回家", "Topic 4: _文章_版友_建議_謝謝", "Topic 5: 離婚_ _結婚_外遇_交往" ] } ], "layout": { "height": 600, "hoverlabel": { "bgcolor": "white", "font": { "family": "Rockwell", "size": 16 } }, "template": { "data": { "bar": [ { "error_x": { "color": "rgb(36,36,36)" }, "error_y": { "color": "rgb(36,36,36)" }, "marker": { "line": { "color": "white", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "bar" } ], "barpolar": [ { "marker": { "line": { "color": "white", "width": 0.5 }, "pattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 } }, "type": "barpolar" } ], "carpet": [ { "aaxis": { "endlinecolor": "rgb(36,36,36)", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "rgb(36,36,36)" }, "baxis": { "endlinecolor": "rgb(36,36,36)", "gridcolor": "white", "linecolor": "white", "minorgridcolor": "white", "startlinecolor": "rgb(36,36,36)" }, "type": "carpet" } ], "choropleth": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "choropleth" } ], "contour": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "contour" } ], "contourcarpet": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "contourcarpet" } ], "heatmap": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "heatmap" } ], "histogram": [ { "marker": { "line": { "color": "white", "width": 0.6 } }, "type": "histogram" } ], "histogram2d": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "histogram2d" } ], "histogram2dcontour": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "histogram2dcontour" } ], "mesh3d": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "type": "mesh3d" } ], "parcoords": [ { "line": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "parcoords" } ], "pie": [ { "automargin": true, "type": "pie" } ], "scatter": [ { "fillpattern": { "fillmode": "overlay", "size": 10, "solidity": 0.2 }, "type": "scatter" } ], "scatter3d": [ { "line": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatter3d" } ], "scattercarpet": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattercarpet" } ], "scattergeo": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattergeo" } ], "scattergl": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattergl" } ], "scattermap": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattermap" } ], "scattermapbox": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scattermapbox" } ], "scatterpolar": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterpolar" } ], "scatterpolargl": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterpolargl" } ], "scatterternary": [ { "marker": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "type": "scatterternary" } ], "surface": [ { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" }, "colorscale": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "type": "surface" } ], "table": [ { "cells": { "fill": { "color": "rgb(237,237,237)" }, "line": { "color": "white" } }, "header": { "fill": { "color": "rgb(217,217,217)" }, "line": { "color": "white" } }, "type": "table" } ] }, "layout": { "annotationdefaults": { "arrowhead": 0, "arrowwidth": 1 }, "autotypenumbers": "strict", "coloraxis": { "colorbar": { "outlinewidth": 1, "tickcolor": "rgb(36,36,36)", "ticks": "outside" } }, "colorscale": { "diverging": [ [ 0, "rgb(103,0,31)" ], [ 0.1, "rgb(178,24,43)" ], [ 0.2, "rgb(214,96,77)" ], [ 0.3, "rgb(244,165,130)" ], [ 0.4, "rgb(253,219,199)" ], [ 0.5, "rgb(247,247,247)" ], [ 0.6, "rgb(209,229,240)" ], [ 0.7, "rgb(146,197,222)" ], [ 0.8, "rgb(67,147,195)" ], [ 0.9, "rgb(33,102,172)" ], [ 1, "rgb(5,48,97)" ] ], "sequential": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ], "sequentialminus": [ [ 0, "#440154" ], [ 0.1111111111111111, "#482878" ], [ 0.2222222222222222, "#3e4989" ], [ 0.3333333333333333, "#31688e" ], [ 0.4444444444444444, "#26828e" ], [ 0.5555555555555556, "#1f9e89" ], [ 0.6666666666666666, "#35b779" ], [ 0.7777777777777778, "#6ece58" ], [ 0.8888888888888888, "#b5de2b" ], [ 1, "#fde725" ] ] }, "colorway": [ "#1F77B4", "#FF7F0E", "#2CA02C", "#D62728", "#9467BD", "#8C564B", "#E377C2", "#7F7F7F", "#BCBD22", "#17BECF" ], "font": { "color": "rgb(36,36,36)" }, "geo": { "bgcolor": "white", "lakecolor": "white", "landcolor": "white", "showlakes": true, "showland": true, "subunitcolor": "white" }, "hoverlabel": { "align": "left" }, "hovermode": "closest", "mapbox": { "style": "light" }, "paper_bgcolor": "white", "plot_bgcolor": "white", "polar": { "angularaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "bgcolor": "white", "radialaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" } }, "scene": { "xaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "yaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "zaxis": { "backgroundcolor": "white", "gridcolor": "rgb(232,232,232)", "gridwidth": 2, "linecolor": "rgb(36,36,36)", "showbackground": true, "showgrid": false, "showline": true, "ticks": "outside", "zeroline": false, "zerolinecolor": "rgb(36,36,36)" } }, "shapedefaults": { "fillcolor": "black", "line": { "width": 0 }, "opacity": 0.3 }, "ternary": { "aaxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "baxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" }, "bgcolor": "white", "caxis": { "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside" } }, "title": { "x": 0.05 }, "xaxis": { "automargin": true, "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside", "title": { "standoff": 15 }, "zeroline": false, "zerolinecolor": "rgb(36,36,36)" }, "yaxis": { "automargin": true, "gridcolor": "rgb(232,232,232)", "linecolor": "rgb(36,36,36)", "showgrid": false, "showline": true, "ticks": "outside", "title": { "standoff": 15 }, "zeroline": false, "zerolinecolor": "rgb(36,36,36)" } } }, "title": { "font": { "color": "Black", "size": 22 }, "text": "Topic Probability Distribution", "x": 0.5, "xanchor": "center", "y": 0.95, "yanchor": "top" }, "width": 800, "xaxis": { "title": { "text": "Probability" } } } } }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 以第1777個文件為例，列出這份文件對每個主題的機率分布\n", "zh_topic_model.visualize_distribution(topic_distr[1777])" ] }, { "cell_type": "markdown", "id": "295f3c21", "metadata": {}, "source": [ "我們以資料集中的其中一篇文章作為範例，查看每個主題的機率分布，其中與情緒詞主題相關的詞彙是最多的，再來是與婚姻關係相關的兩個主題的詞彙較多，討論到小孩和家庭生活的詞彙則較少。" ] }, { "cell_type": "code", "execution_count": null, "id": "79800333", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "[('房子', 0.08832980862301461),\n", " (' ', 0.08206604884387665),\n", " ('房貸', 0.05096740115727702),\n", " ('萬', 0.047356587645044315),\n", " ('錢', 0.0450103800079604),\n", " ('薪水', 0.04135365237991898),\n", " ('財產', 0.03950471267694048),\n", " ('買', 0.03455065077692002),\n", " ('工作', 0.03227272405986625),\n", " ('貸款', 0.03135481554556126),\n", " ('費用', 0.02734955195859806),\n", " ('買房', 0.02721027073860453),\n", " ('收入', 0.024952198329710087),\n", " ('存款', 0.02376174368712919),\n", " ('名下', 0.02236766307496004),\n", " ('小孩', 0.022094595487329578),\n", " ('開銷', 0.021967702675287745),\n", " ('負擔', 0.01928951669898701),\n", " ('住', 0.018815296571449015),\n", " ('一半', 0.018162093209078425),\n", " ('老公', 0.017709298496047788),\n", " ('老婆', 0.01752294289544913),\n", " ('薪資', 0.017494089796292113),\n", " ('投資', 0.01704722423177808),\n", " ('一個月', 0.01670727403634324),\n", " ('付', 0.016667227857624485),\n", " ('賺', 0.01634891501423118),\n", " ('年薪', 0.01633017406466438),\n", " ('家用', 0.016122946393990072),\n", " ('說', 0.015515271195159285)]" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 列出主題的代表詞和其對應的權重\n", "zh_topic_model.get_topic(2)" ] }, { "cell_type": "markdown", "id": "1ab7f9e8", "metadata": {}, "source": [ "我們查看了與房屋、財產相關的這個主題的代表字及對應的權重，主要有房子、房貸、錢、薪水、工作、貸款等，主要都圍繞在房屋投資及工作收入，與當今社會對婚姻狀況討論最多的議題相關。" ] }, { "cell_type": "markdown", "id": "1fa4f4bd", "metadata": {}, "source": [ "查看特定文章的主題分佈" ] }, { "cell_type": "code", "execution_count": null, "id": "2c83e5fd", "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 視覺化顯示主題-詞彙分佈\n", "topic_n = 2\n", "data = zh_topic_model.get_topic(topic_n)\n", "\n", "# 轉換為DataFrame\n", "df = pd.DataFrame(data, columns=['word', 'prob'])\n", "df = df[df['word'] != ' ']\n", "\n", "# 根據prob排序並選出前10名\n", "top_10 = df.sort_values('prob', ascending=False).head(10)\n", "\n", "# 畫出長條圖\n", "plt.figure(figsize=(10,6))\n", "plt.barh(top_10['word'], top_10['prob'], color='navy')\n", "plt.xlabel('機率')\n", "plt.title(f'主題 {topic_n} 詞彙機率前10名')\n", "plt.gca().invert_yaxis()\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "6733a2b3", "metadata": {}, "source": [ "我們也列出了前十高的詞彙，從分布長條圖能看出，房子是提及最多的，其次是房貸，還有如萬、錢這類與金錢相關的用字，說明這個主題中討論最多的還是多圍繞在房產及金錢的劃分，可能可以反映出婚姻關係中與房屋相關的經濟負擔、收入財產分配的爭議是最常見的。" ] }, { "cell_type": "code", "execution_count": null, "id": "f815a762", "metadata": {}, "outputs": [], "source": [ "from bertopic.representation import KeyBERTInspired\n", "# KeyBERT\n", "keybert = KeyBERTInspired()\n", "\n", "# 設定HDBscan模型\n", "hdbscan_model = HDBSCAN(min_cluster_size=5, min_samples=30)\n", "\n", "# 定義我們要用到的representation model\n", "representation_model = {\n", " \"KeyBERT\": keybert,\n", "}" ] }, { "cell_type": "code", "execution_count": null, "id": "91e02591", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "2025-05-10 03:41:55,606 - BERTopic - Dimensionality - Fitting the dimensionality reduction algorithm\n", "2025-05-10 03:43:13,490 - BERTopic - Dimensionality - Completed ✓\n", "2025-05-10 03:43:13,492 - BERTopic - Cluster - Start clustering the reduced embeddings\n", "2025-05-10 03:43:19,574 - BERTopic - Cluster - Completed ✓\n", "2025-05-10 03:43:19,586 - BERTopic - Representation - Fine-tuning topics using representation models.\n", "2025-05-10 03:43:32,595 - BERTopic - Representation - Completed ✓\n" ] } ], "source": [ "# 建立BERTopic模型\n", "representation_topic_model = BERTopic(\n", " # Sub-models\n", " embedding_model=bert_sentence_model,\n", " vectorizer_model=jieba_vectorizer,\n", " # 設定Representation model\n", " representation_model=representation_model,\n", " # Hyperparameters\n", " top_n_words=30,\n", " verbose=True\n", ")\n", "\n", "# Train model\n", "topics, probs = representation_topic_model.fit_transform(docs_zh, all_embeddings) " ] }, { "cell_type": "code", "execution_count": null, "id": "cf202515", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
TopicCountNameRepresentationKeyBERTRepresentative_Docs
0-178153-1_說_小孩_妳_[說, 小孩, 妳, , 老婆, 想, 老公, 真的, 做, 工作, 離婚, 孩子, 生活...[有沒有, 好好, 回家, 我媽, 房子, 希望, 開心, 婆家, 幫忙, 孩子][當她想吃東西時, 所以原又有回公婆家, 無意中從爸爸手機中的訊息內容發現跟一個女生聊天所以...
1018520_ ___[ , , , , , , , , , , , , , , , , , , , , , , ...[, , , , , , , , , ][, , ]
218831_洗_衣服_地板_乾淨[洗, 衣服, 地板, 乾淨, 煮, 洗碗, 冰箱, 尿布, 浴室, 消毒, 空間, 洗碗機...[洗碗機, 曬衣服, 洗衣機, 洗衣服, 晾衣服, 洗手, 拖地, 吸塵器, 機器人, 洗碗][我常常跟她說洗不乾淨, 現在兩人衣服也都我在洗, 只有洗自己的衣服]
326812_老婆_老公_岳父_你媽[老婆, 老公, 岳父, 你媽, 婆婆, 公婆, 喊停, 老爺, 不爽, 媳婦, 老馬, 喊...[老姑婆, 你媽, 我愛我, 噴你媽, 老娘, 你家, 老公, 掛你, 玩不動, 老婆][他老婆不可以, 所以才處處覺得老婆的不是, 你老婆根本有問題]
434893_個性_脾氣_性格_很會[個性, 脾氣, 性格, 很會, 喜歡, 善良, 冷淡, 強勢, 溫和, 很強, 衛生習慣,...[個性, 性格, 大男人主義, 好脾氣, 衛生習慣, 口頭禪, 和藹可親, 很強, 沒什麼,...[這樣的個性, 對上這種個性的人, 同時她的個性也很好]
.....................
65865710657_迴圈_無窮_循環_無限[迴圈, 無窮, 循環, 無限, 沒學, 情況, 反感, 後援, 差異, 反正, , , ...[迴圈, 無限, 無窮, 循環, 情況, 後援, 沒學, 反正, 差異, 反感][無限迴圈, 反正無窮迴圈, 無窮迴圈]
65965810658_偷不著_如來_偷_一把[偷不著, 如來, 偷, 一把, 拚, 有趣, 不給, 玩, 做, , , , , , ,...[偷不著, 不給, 有趣, 偷, 一把, 拚, 玩, 如來, 做, ][不如拚一把, 不如來的有趣, 偷不如偷不著]
66065910659_做得還_盡善盡美_面面俱到_幅度[做得還, 盡善盡美, 面面俱到, 幅度, 自食其力, 有作, 太小, 到位, 承認, 終究...[面面俱到, 盡善盡美, 做得還, 到位, 打算, 真的, 自食其力, 幅度, 我會, 算是][我自己很多地方有作到位, 我認為她進步幅度太小, 短期間內真的是很難做到盡善盡美]
66166010660_韓劇_電影_裡看_整部[韓劇, 電影, 裡看, 整部, 聽相聲, 聽陸劇, 過劇, 有累, 偶然, 動畫, 看劇,...[聽陸劇, 看電視, 聽相聲, 小時候, 休息時間, 偶然, 抬頭, 天天, 看劇, 那天][而且兩次就把整部韓劇看完, 而且說實在的整天晾在家裡看韓劇休息, 我跟老公會一起看韓劇電影動畫]
66266110661_人身攻擊_論點_言論自由_言論[人身攻擊, 論點, 言論自由, 言論, 攻擊, 恰當, 酸言酸語, 片面, 指出, 無辜,...[言論自由, 人身攻擊, 酸言酸語, 尊重, 理性, 攻擊, 恰當, 反對, 指出, 傷害][只要我感受到人身攻擊或是不受尊重, 不過很多人對人不對事對我的人身攻擊, 我沒有人身攻擊]
\n", "

663 rows × 6 columns

\n", "
" ], "text/plain": [ " Topic Count Name \\\n", "0 -1 78153 -1_說_小孩_妳_ \n", "1 0 1852 0_ ___ \n", "2 1 883 1_洗_衣服_地板_乾淨 \n", "3 2 681 2_老婆_老公_岳父_你媽 \n", "4 3 489 3_個性_脾氣_性格_很會 \n", ".. ... ... ... \n", "658 657 10 657_迴圈_無窮_循環_無限 \n", "659 658 10 658_偷不著_如來_偷_一把 \n", "660 659 10 659_做得還_盡善盡美_面面俱到_幅度 \n", "661 660 10 660_韓劇_電影_裡看_整部 \n", "662 661 10 661_人身攻擊_論點_言論自由_言論 \n", "\n", " Representation \\\n", "0 [說, 小孩, 妳, , 老婆, 想, 老公, 真的, 做, 工作, 離婚, 孩子, 生活... \n", "1 [ , , , , , , , , , , , , , , , , , , , , , , ... \n", "2 [洗, 衣服, 地板, 乾淨, 煮, 洗碗, 冰箱, 尿布, 浴室, 消毒, 空間, 洗碗機... \n", "3 [老婆, 老公, 岳父, 你媽, 婆婆, 公婆, 喊停, 老爺, 不爽, 媳婦, 老馬, 喊... \n", "4 [個性, 脾氣, 性格, 很會, 喜歡, 善良, 冷淡, 強勢, 溫和, 很強, 衛生習慣,... \n", ".. ... \n", "658 [迴圈, 無窮, 循環, 無限, 沒學, 情況, 反感, 後援, 差異, 反正, , , ... \n", "659 [偷不著, 如來, 偷, 一把, 拚, 有趣, 不給, 玩, 做, , , , , , ,... \n", "660 [做得還, 盡善盡美, 面面俱到, 幅度, 自食其力, 有作, 太小, 到位, 承認, 終究... \n", "661 [韓劇, 電影, 裡看, 整部, 聽相聲, 聽陸劇, 過劇, 有累, 偶然, 動畫, 看劇,... \n", "662 [人身攻擊, 論點, 言論自由, 言論, 攻擊, 恰當, 酸言酸語, 片面, 指出, 無辜,... \n", "\n", " KeyBERT \\\n", "0 [有沒有, 好好, 回家, 我媽, 房子, 希望, 開心, 婆家, 幫忙, 孩子] \n", "1 [, , , , , , , , , ] \n", "2 [洗碗機, 曬衣服, 洗衣機, 洗衣服, 晾衣服, 洗手, 拖地, 吸塵器, 機器人, 洗碗] \n", "3 [老姑婆, 你媽, 我愛我, 噴你媽, 老娘, 你家, 老公, 掛你, 玩不動, 老婆] \n", "4 [個性, 性格, 大男人主義, 好脾氣, 衛生習慣, 口頭禪, 和藹可親, 很強, 沒什麼,... \n", ".. ... \n", "658 [迴圈, 無限, 無窮, 循環, 情況, 後援, 沒學, 反正, 差異, 反感] \n", "659 [偷不著, 不給, 有趣, 偷, 一把, 拚, 玩, 如來, 做, ] \n", "660 [面面俱到, 盡善盡美, 做得還, 到位, 打算, 真的, 自食其力, 幅度, 我會, 算是] \n", "661 [聽陸劇, 看電視, 聽相聲, 小時候, 休息時間, 偶然, 抬頭, 天天, 看劇, 那天] \n", "662 [言論自由, 人身攻擊, 酸言酸語, 尊重, 理性, 攻擊, 恰當, 反對, 指出, 傷害] \n", "\n", " Representative_Docs \n", "0 [當她想吃東西時, 所以原又有回公婆家, 無意中從爸爸手機中的訊息內容發現跟一個女生聊天所以... \n", "1 [, , ] \n", "2 [我常常跟她說洗不乾淨, 現在兩人衣服也都我在洗, 只有洗自己的衣服] \n", "3 [他老婆不可以, 所以才處處覺得老婆的不是, 你老婆根本有問題] \n", "4 [這樣的個性, 對上這種個性的人, 同時她的個性也很好] \n", ".. ... \n", "658 [無限迴圈, 反正無窮迴圈, 無窮迴圈] \n", "659 [不如拚一把, 不如來的有趣, 偷不如偷不著] \n", "660 [我自己很多地方有作到位, 我認為她進步幅度太小, 短期間內真的是很難做到盡善盡美] \n", "661 [而且兩次就把整部韓劇看完, 而且說實在的整天晾在家裡看韓劇休息, 我跟老公會一起看韓劇電影動畫] \n", "662 [只要我感受到人身攻擊或是不受尊重, 不過很多人對人不對事對我的人身攻擊, 我沒有人身攻擊] \n", "\n", "[663 rows x 6 columns]" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# 觀察KeyBERT微調後的主題表示\n", "representation_topic_model.get_topic_info()" ] }, { "cell_type": "markdown", "id": "dddf4700", "metadata": {}, "source": [ "經過微調後，主題數量從原本的六個大幅增加到六百多個，顯示模型在嘗試更細緻地劃分，但可能原本文章的內容便差異不大，導致在過度細緻的劃分下，被迫分成許多小主題，雖然這可以捕捉到更多具體的婚姻議題，如家事分工、親子互動、娛樂活動等，但這種過度細化的結果可能反而不利於整體議題的統整。" ] } ], "metadata": { "accelerator": "GPU", "colab": { "gpuType": "T4", "provenance": [] }, "kernelspec": { "display_name": "Syllabus", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.11" } }, "nbformat": 4, "nbformat_minor": 5 }