{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Tf-idf Vectorizer, Logistic Regression" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Selected Category: description\n", "description has 280 samples;\n", "installation has 70 samples;\n", "invocation has 70 samples;\n", "citation has 70 samples;\n", "Selected Category: installation\n", "description has 200 samples;\n", "installation has 800 samples;\n", "invocation has 200 samples;\n", "citation has 200 samples;\n", "Selected Category: invocation\n", "description has 279 samples;\n", "installation has 279 samples;\n", "invocation has 1118 samples;\n", "citation has 279 samples;\n", "Selected Category: citation\n", "description has 77 samples;\n", "installation has 77 samples;\n", "invocation has 77 samples;\n", "citation has 309 samples;\n", "{'description': excerpt description\n", "0 Puppeteer is a Node library which provides a h... True\n", "1 The major contributors of this repository incl... True\n", "2 Integral Regression is initially described in ... True\n", "3 We build a 3D pose estimation system based mai... True\n", "4 The Integral Regression is also known as soft-... True\n", "5 This is an official implementation for Integra... True\n", "6 The original implementation is based on our in... True\n", "7 LibGEOS is a LGPL-licensed package for manipul... True\n", "8 Among other things, it allows you to parse Wel... True\n", "9 This repository contains the experiments in th... True\n", "10 For the results presented in the paper, we did... True\n", "11 Batch normalization is currently not supported... True\n", "12 Open-source Ground Penetrating Radar processin... True\n", "13 Pytorch implementation for high-resolution (e.... True\n", "14 The PVGeo Python package contains VTK powered ... True\n", "15 A PyVista (and VTK) interface for the Open Min... True\n", "16 GeoNotebook is an application that provides cl... True\n", "17 Fiona is OGR's neat and nimble API for Python ... True\n", "18 Fiona is designed to be simple and dependable.... True\n", "19 Shapely is a BSD-licensed Python package for m... True\n", "20 Rain streaks can severely degrade the visibili... True\n", "21 The pytorch branch contains: True\n", "22 the pytorch implementation of Peak Response Ma... True\n", "23 the PASCAL-VOC demo (training, inference, and ... True\n", "24 Lithology and stratigraphic logs for wells and... True\n", "25 This Python module allows you to: True\n", "26 Interactively control an instance of ANSYS v14... True\n", "27 Extract data directly from binary ANSYS v14.5+... True\n", "28 Rapidly read in binary result (.rst), binary m... True\n", "29 Official implementation of GANimation. In this... True\n", ".. ... ...\n", "460 year={2018} False\n", "461 } False\n", "462 pages={262--277}, False\n", "463 @InProceedings{Lim_2017_CVPR_Workshops, False\n", "464 author = {Lars Mescheder and Sebastian Nowozin... False\n", "465 year = {2018} False\n", "466 Dieter Werthmüller, prisae False\n", "467 Citation False\n", "468 @inproceedings{DeepMVS, False\n", "469 Matteo Ravasi, mrava87 False\n", "470 year={2018} False\n", "471 Key Laboratory of Machine Perception, Shenzhen... False\n", "472 } False\n", "473 } False\n", "474 Author = {Xizhou Zhu, Yujie Wang, Jifeng Dai, ... False\n", "475 @inproceedings{tang2018quantized, False\n", "476 [Paper Link] (CVPR'18) False\n", "477 Title = {Flow-Guided Feature Aggregation for V... False\n", "478 booktitle = {IEEE Conferene on Computer Vision... False\n", "479 Year = {2016} False\n", "480 booktitle = \"Conference on Computer Vision and... False\n", "481 pages = {1450}, False\n", "482 Yu, (2018). PyGeoPressure: Geopressure Predict... False\n", "483 year = {2018}, False\n", "484 year = {2018} False\n", "485 Conference = {ICCV}, False\n", "486 title={Scale-recurrent Network for Deep Image ... False\n", "487 BibTex: False\n", "488 Key Laboratory of Machine Perception (MOE), Sc... False\n", "489 title={Image Generation from Scene Graphs}, False\n", "\n", "[490 rows x 2 columns], 'installation': excerpt installation\n", "0 ocker is an operating-system-level-visualizati... False\n", "1 TensorFlow is an open source software library ... False\n", "2 Puppeteer is a Node library which provides a h... False\n", "3 The Laplacian Pyramid Super-Resolution Network... False\n", "4 Segyio is a small LGPL licensed C library for ... False\n", "5 GeoNotebook is an application that provides cl... False\n", "6 RetinaNet False\n", "7 This repository only contains the core compone... False\n", "8 Hankel transforms (wavenumber-frequency to spa... False\n", "9 Below we show some example scene graphs along ... False\n", "10 All planar measurements are expected to follow... False\n", "11 Currently, segyio supports: False\n", "12 construction and interactive editing of spatia... False\n", "13 Declarative: React makes it painless to create... False\n", "14 Luckily, many iterative methods (e.g. cg, lsqr... False\n", "15 A highly efficient JavaScript library for slic... False\n", "16 We proposed to evaluate the detection accuracy... False\n", "17 If you use our codes or datasets in your work,... False\n", "18 This code was made public to share our researc... False\n", "19 It now takes a period of time closer to 24 hou... False\n", "20 In addition to applications in teaching and re... False\n", "21 dipole: infinitesimal small dipoles oriented a... False\n", "22 Rapidly read in binary result (.rst), binary m... False\n", "23 Theano allows the automated computation of gra... False\n", "24 Intuitive plotting routines with matplotlib si... False\n", "25 A Python package for pore pressure prediction ... False\n", "26 The pytorch branch contains: False\n", "27 Additional backbone architectures may be easil... False\n", "28 Introduction False\n", "29 This code is written in Chainer. For PyTorch u... False\n", "... ... ...\n", "1370 @article{sun2018integral, False\n", "1371 booktitle={arXiv}, False\n", "1372 title = {Enhanced Deep Residual Networks for S... False\n", "1373 Booktitle = {European Conference on Compu... False\n", "1374 author = {Xinlei Chen and Li-Jia Li and Li Fei... False\n", "1375 } False\n", "1376 year = {2018} False\n", "1377 year={2018} False\n", "1378 Xin Tao, Hongyun Gao, Xiaoyong Shen, Jue Wang,... False\n", "1379 CVPR 2018 False\n", "1380 False\n", "1381 year = {2018} False\n", "1382 booktitle = {CVPR}, False\n", "1383 Author = {Xizhou Zhu, Yujie Wang, Jifeng Dai, ... False\n", "1384 Dieter Werthmüller, prisae False\n", "1385 This software is based on ideas published ther... False\n", "1386 @INPROCEEDINGS{Mescheder2018ICML, False\n", "1387 Title = {Flow-Guided Feature Aggregation for V... False\n", "1388 @inproceedings{tao2018srndeblur, False\n", "1389 If you find the code and datasets useful in yo... False\n", "1390 de la Varga, M., Schaaf, A., and Wellmann, F.:... False\n", "1391 @inproceedings{tang2018quantized, False\n", "1392 @InProceedings{kato2018renderer False\n", "1393 All releases have a Zenodo-DOI, provided on th... False\n", "1394 journal = {Journal of Open Source Software} False\n", "1395 journal={arXiv preprint arXiv:1809.06079}, False\n", "1396 year={2018} False\n", "1397 Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhon... False\n", "1398 } False\n", "1399 Huikai Wu, Shuai Zheng, Junge Zhang, Kaiqi Huang False\n", "\n", "[1400 rows x 2 columns], 'invocation': excerpt invocation\n", "0 This repository is implemented by Yuqing Zhu, ... False\n", "1 Python/Cython wrapper of Marco Attene's wonder... False\n", "2 This baseline is run on dbnet-2018 challenge d... False\n", "3 All planar measurements are expected to follow... False\n", "4 If you give it a collection of years of tweet ... False\n", "5 Introduction False\n", "6 In addition to applications in teaching and re... False\n", "7 Quadrature with extrapolation QWE False\n", "8 Introduction False\n", "9 Basically, he wears a top hat, lives in your c... False\n", "10 The input is assumed to represent a single clo... False\n", "11 Pytorch implementation for high-resolution (e.... False\n", "12 This project aims to automate the manual proce... False\n", "13 Learn Once, Write Anywhere: We don't make assu... False\n", "14 mplleaflet is a Python library that converts a... False\n", "15 GeoNotebook is an application that provides cl... False\n", "16 Faster R-CNN False\n", "17 A Jupyter / Leaflet bridge enabling interactiv... False\n", "18 This is the code for the paper False\n", "19 FGFA is end-to-end trainable for the task of v... False\n", "20 RPN False\n", "21 Very lite but extendable mapping framework to ... False\n", "22 Eaton's method and Parameter Optimization False\n", "23 VGG16 False\n", "24 We build a 3D pose estimation system based mai... False\n", "25 PyVista is a helper module for the Visualizati... False\n", "26 Introduction False\n", "27 project loading False\n", "28 Fiona is designed to be simple and dependable.... False\n", "29 RetinaNet False\n", "... ... ...\n", "1925 @inproceedings{chen2018domain, False\n", "1926 @article{yu2018pygeopressure, False\n", "1927 title = {{PyGeoPressure}: {Geopressure} {Predi... False\n", "1928 pages = {922} False\n", "1929 Ting-Chun Wang1, Ming-Yu Liu1, Jun-Yan Zhu2, G... False\n", "1930 url = {https://doi.org/10.21105/joss.01450}, False\n", "1931 title={Domain Adaptive Faster R-CNN for Object... False\n", "1932 pages={6546--6555}, False\n", "1933 title={Recurrent Squeeze-and-Excitation Contex... False\n", "1934 booktitle = {IEEE Conference on Computer Visio... False\n", "1935 Matteo Ravasi, mrava87 False\n", "1936 } False\n", "1937 year={2018} False\n", "1938 booktitle={ECCV}, False\n", "1939 month = {may}, False\n", "1940 All releases have a Zenodo-DOI, provided on th... False\n", "1941 @INPROCEEDINGS{Mescheder2018ICML, False\n", "1942 Video-to-Video Synthesis False\n", "1943 title={An Integral Pose Regression System for ... False\n", "1944 booktitle={Computer Vision and Pattern Regogni... False\n", "1945 year = {2018} False\n", "1946 title = {Spatial Memory for Context Reasoning ... False\n", "1947 year={2017} False\n", "1948 Proceedings of the IEEE Conference on Computer... False\n", "1949 Tristan van Leeuwen, TristanvanLeeuwen False\n", "1950 } False\n", "1951 booktitle={CVPR}, False\n", "1952 year={2018}, False\n", "1953 title={CU-Net: Coupled U-Nets}, False\n", "1954 } False\n", "\n", "[1955 rows x 2 columns], 'citation': excerpt citation\n", "0 Below we show some example scene graphs along ... False\n", "1 Faster, Better and Lighter for image processin... False\n", "2 To reproduce the quantitative results shown in... False\n", "3 Bowers' method and Parameter Optimization False\n", "4 Tilematrix supports metatiling and tile buffer... False\n", "5 Direct access to mesh analysis and transformat... False\n", "6 Luckily, many iterative methods (e.g. cg, lsqr... False\n", "7 the PASCAL-VOC demo (training, inference, and ... False\n", "8 gpr: calculates the ground-penetrating radar r... False\n", "9 A highly efficient JavaScript library for slic... False\n", "10 Intuitive plotting routines with matplotlib si... False\n", "11 Each branch in the git repository corresponds ... False\n", "12 Fast R-CNN False\n", "13 SEG-Y Revisions False\n", "14 If you give it all of OpenStreetMap and zoom o... False\n", "15 Lithology and stratigraphic logs for wells and... False\n", "16 personal website + blog for every github user False\n", "17 Note this is not a package for reading LiDAR d... False\n", "18 This is a yeoman generator for ArcGIS API for ... False\n", "19 For simplicity, each dot represents one U-Net.... False\n", "20 The electromagnetic modeller empymod can model... False\n", "21 Remote Geomod: From GoogleEarth to 3-D Geology False\n", "22 The mapshaper command line program supports es... False\n", "23 PySAL, the Python spatial analysis library, is... False\n", "24 For the results presented in the paper, we did... False\n", "25 Very lite but extendable mapping framework to ... False\n", "26 A Python package for pore pressure prediction ... False\n", "27 VGG16 False\n", "28 This is a Python 2.7 and 3.3+ package to read ... False\n", "29 We proposed to evaluate the detection accuracy... False\n", ".. ... ...\n", "510 booktitle = {Computer Vision and Pattern Recog... True\n", "511 year={2018} True\n", "512 } True\n", "513 Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhon... True\n", "514 Yulun Zhang, Yapeng Tian, Yu Kong, Bineng Zhon... True\n", "515 @InProceedings{Lim_2017_CVPR_Workshops, True\n", "516 author = {Lim, Bee and Son, Sanghyun and Kim, ... True\n", "517 title = {Enhanced Deep Residual Networks for S... True\n", "518 booktitle = {The IEEE Conference on Computer V... True\n", "519 month = {July}, True\n", "520 year = {2017} True\n", "521 } True\n", "522 @inproceedings{zhang2018residual, True\n", "523 title={Residual Dense Network for Image Super-... True\n", "524 author={Zhang, Yulun and Tian, Yapeng and Kong... True\n", "525 booktitle={CVPR}, True\n", "526 year={2018} True\n", "527 @article{zhang2018rdnir, True\n", "528 title={Residual Dense Network for Image Restor... True\n", "529 booktitle={arXiv}, True\n", "530 @inproceedings{tang2018quantized, True\n", "531 title={Quantized densely connected U-Nets for ... True\n", "532 author={Tang, Zhiqiang and Peng, Xi and Geng, ... True\n", "533 booktitle={ECCV}, True\n", "534 year={2018} True\n", "535 } True\n", "536 @inproceedings{tang2018cu, True\n", "537 title={CU-Net: Coupled U-Nets}, True\n", "538 author={Tang, Zhiqiang and Peng, Xi and Geng, ... True\n", "539 booktitle={BMVC}, True\n", "\n", "[540 rows x 2 columns]}\n" ] } ], "source": [ "from setup_corpus import build_corpora\n", "corpora = build_corpora()\n", "print(corpora)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "doing something haha" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "from sklearn.model_selection import cross_val_score, cross_validate, StratifiedKFold\n", "from sklearn.feature_extraction.text import TfidfVectorizer\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.pipeline import make_pipeline\n", "from sklearn.metrics import confusion_matrix, classification_report, accuracy_score, roc_curve, auc, precision_recall_curve, average_precision_score\n", "pipeline = make_pipeline(TfidfVectorizer(), LogisticRegression(solver='liblinear'))\n", "\n", "cv = StratifiedKFold(n_splits = 5, shuffle=True)" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Category: description\n", "Scores: [0.74489796 0.89795918 0.85714286 0.87755102 0.80612245]\n", "Accuracy: 0.8367 (+/- 0.1103)\n", "Category: installation\n", "Scores: [0.925 0.90357143 0.88214286 0.925 0.87142857]\n", "Accuracy: 0.9014 (+/- 0.0437)\n", "Category: invocation\n", "Scores: [0.87244898 0.81632653 0.84910486 0.86923077 0.85384615]\n", "Accuracy: 0.8522 (+/- 0.0400)\n", "Category: citation\n", "Scores: [0.9266055 0.84259259 0.87037037 0.9537037 0.85981308]\n", "Accuracy: 0.8906 (+/- 0.0846)\n" ] } ], "source": [ "for category in corpora:\n", " scores = cross_val_score(pipeline, corpora[category].excerpt, corpora[category][category], cv=cv)\n", " print(f\"Category: {category}\\nScores: {scores}\\nAccuracy: {scores.mean():.4f} (+/- {scores.std()*2:.4f})\")" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [], "source": [ "import numpy as np\n", "from scipy import interp\n", "import matplotlib.pyplot as plt" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Description ROC\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "X = corpora['description'].excerpt\n", "y = corpora['description'].description\n", "tprs = []\n", "aucs = []\n", "mean_fpr = np.linspace(0, 1, 100)\n", " \n", "i = 0\n", "print('Description ROC')\n", "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " # Compute ROC curve and area under the curve\n", " fpr, tpr, thresholds = roc_curve(y[test], probas_[:, 1])\n", " tprs.append(interp(mean_fpr, fpr, tpr))\n", " tprs[-1][0] = 0.0\n", " roc_auc = auc(fpr, tpr)\n", " aucs.append(roc_auc)\n", " plt.plot(fpr, tpr, lw=1, alpha=0.3, label='ROC fold %d (AUC = %0.2f)' % (i, roc_auc))\n", " i+=1\n", "plt.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r',\n", " label='Chance', alpha=.8)\n", "\n", "mean_tpr = np.mean(tprs, axis=0)\n", "mean_tpr[-1] = 1.0\n", "mean_auc = auc(mean_fpr, mean_tpr)\n", "std_auc = np.std(aucs)\n", "plt.plot(mean_fpr, mean_tpr, color='b',\n", " label=r'Mean ROC (AUC = %0.2f $\\pm$ %0.2f)' % (mean_auc, std_auc),\n", " lw=2, alpha=.8)\n", "\n", "std_tpr = np.std(tprs, axis=0)\n", "tprs_upper = np.minimum(mean_tpr + std_tpr, 1)\n", "tprs_lower = np.maximum(mean_tpr - std_tpr, 0)\n", "plt.fill_between(mean_fpr, tprs_lower, tprs_upper, color='grey', alpha=.2,\n", " label=r'$\\pm$ 1 std. dev.')\n", "\n", "plt.xlim([-0.05, 1.05])\n", "plt.ylim([-0.05, 1.05])\n", "plt.xlabel('False Positive Rate')\n", "plt.ylabel('True Positive Rate')\n", "plt.title('ROC Curve for Description Classification')\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " precision, recall, _ = precision_recall_curve(y[test], probas_[:,1])\n", "\n", " plt.step(recall, precision, alpha=0.2,\n", " where='post', label=f'average precision={average_precision_score(y[test], probas_[:,1])}')\n", "\n", " plt.xlabel('Recall')\n", " plt.ylabel('Precision')\n", " plt.ylim([0.0, 1.05])\n", " plt.xlim([0.0, 1.0])\n", "plt.title('Description Precision-Recall curve'.format(\n", " average_precision_score(y[test], probas_[:,1])))\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 7, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Installation ROC\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "X = corpora['installation'].excerpt\n", "y = corpora['installation'].installation\n", "tprs = []\n", "aucs = []\n", "mean_fpr = np.linspace(0, 1, 100)\n", " \n", "i = 0\n", "print('Installation ROC')\n", "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " # Compute ROC curve and area under the curve\n", " fpr, tpr, thresholds = roc_curve(y[test], probas_[:, 1])\n", " tprs.append(interp(mean_fpr, fpr, tpr))\n", " tprs[-1][0] = 0.0\n", " roc_auc = auc(fpr, tpr)\n", " aucs.append(roc_auc)\n", " plt.plot(fpr, tpr, lw=1, alpha=0.3, label='ROC fold %d (AUC = %0.2f)' % (i, roc_auc))\n", " i+=1\n", "plt.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r',\n", " label='Chance', alpha=.8)\n", "\n", "mean_tpr = np.mean(tprs, axis=0)\n", "mean_tpr[-1] = 1.0\n", "mean_auc = auc(mean_fpr, mean_tpr)\n", "std_auc = np.std(aucs)\n", "plt.plot(mean_fpr, mean_tpr, color='b',\n", " label=r'Mean ROC (AUC = %0.2f $\\pm$ %0.2f)' % (mean_auc, std_auc),\n", " lw=2, alpha=.8)\n", "\n", "std_tpr = np.std(tprs, axis=0)\n", "tprs_upper = np.minimum(mean_tpr + std_tpr, 1)\n", "tprs_lower = np.maximum(mean_tpr - std_tpr, 0)\n", "plt.fill_between(mean_fpr, tprs_lower, tprs_upper, color='grey', alpha=.2,\n", " label=r'$\\pm$ 1 std. dev.')\n", "\n", "plt.xlim([-0.05, 1.05])\n", "plt.ylim([-0.05, 1.05])\n", "plt.xlabel('False Positive Rate')\n", "plt.ylabel('True Positive Rate')\n", "plt.title('ROC Curve for Installation Classification')\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " precision, recall, _ = precision_recall_curve(y[test], probas_[:,1])\n", "\n", " plt.step(recall, precision, alpha=0.2,\n", " where='post', label=f'average precision={average_precision_score(y[test], probas_[:,1])}')\n", "\n", " plt.xlabel('Recall')\n", " plt.ylabel('Precision')\n", " plt.ylim([0.0, 1.05])\n", " plt.xlim([0.0, 1.0])\n", "plt.title('Installation Precision-Recall curve'.format(\n", " average_precision_score(y[test], probas_[:,1])))\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Installation ROC\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "X = corpora['invocation'].excerpt\n", "y = corpora['invocation'].invocation\n", "tprs = []\n", "aucs = []\n", "mean_fpr = np.linspace(0, 1, 100)\n", " \n", "i = 0\n", "print('Installation ROC')\n", "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " # Compute ROC curve and area under the curve\n", " fpr, tpr, thresholds = roc_curve(y[test], probas_[:, 1])\n", " tprs.append(interp(mean_fpr, fpr, tpr))\n", " tprs[-1][0] = 0.0\n", " roc_auc = auc(fpr, tpr)\n", " aucs.append(roc_auc)\n", " plt.plot(fpr, tpr, lw=1, alpha=0.3, label='ROC fold %d (AUC = %0.2f)' % (i, roc_auc))\n", " i+=1\n", "plt.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r',\n", " label='Chance', alpha=.8)\n", "\n", "mean_tpr = np.mean(tprs, axis=0)\n", "mean_tpr[-1] = 1.0\n", "mean_auc = auc(mean_fpr, mean_tpr)\n", "std_auc = np.std(aucs)\n", "plt.plot(mean_fpr, mean_tpr, color='b',\n", " label=r'Mean ROC (AUC = %0.2f $\\pm$ %0.2f)' % (mean_auc, std_auc),\n", " lw=2, alpha=.8)\n", "\n", "std_tpr = np.std(tprs, axis=0)\n", "tprs_upper = np.minimum(mean_tpr + std_tpr, 1)\n", "tprs_lower = np.maximum(mean_tpr - std_tpr, 0)\n", "plt.fill_between(mean_fpr, tprs_lower, tprs_upper, color='grey', alpha=.2,\n", " label=r'$\\pm$ 1 std. dev.')\n", "\n", "plt.xlim([-0.05, 1.05])\n", "plt.ylim([-0.05, 1.05])\n", "plt.xlabel('False Positive Rate')\n", "plt.ylabel('True Positive Rate')\n", "plt.title('ROC Curve for Invocation Classification')\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " precision, recall, _ = precision_recall_curve(y[test], probas_[:,1])\n", "\n", " plt.step(recall, precision, alpha=0.2,\n", " where='post', label=f'average precision={average_precision_score(y[test], probas_[:,1])}')\n", "\n", " plt.xlabel('Recall')\n", " plt.ylabel('Precision')\n", " plt.ylim([0.0, 1.05])\n", " plt.xlim([0.0, 1.0])\n", "plt.title('Invocation Precision-Recall curve'.format(\n", " average_precision_score(y[test], probas_[:,1])))\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Citation ROC\n" ] }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "X = corpora['citation'].excerpt\n", "y = corpora['citation'].citation\n", "tprs = []\n", "aucs = []\n", "mean_fpr = np.linspace(0, 1, 100)\n", " \n", "i = 0\n", "print('Citation ROC')\n", "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " # Compute ROC curve and area under the curve\n", " fpr, tpr, thresholds = roc_curve(y[test], probas_[:, 1])\n", " tprs.append(interp(mean_fpr, fpr, tpr))\n", " tprs[-1][0] = 0.0\n", " roc_auc = auc(fpr, tpr)\n", " aucs.append(roc_auc)\n", " plt.plot(fpr, tpr, lw=1, alpha=0.3, label='ROC fold %d (AUC = %0.2f)' % (i, roc_auc))\n", " i+=1\n", "plt.plot([0, 1], [0, 1], linestyle='--', lw=2, color='r',\n", " label='Chance', alpha=.8)\n", "\n", "mean_tpr = np.mean(tprs, axis=0)\n", "mean_tpr[-1] = 1.0\n", "mean_auc = auc(mean_fpr, mean_tpr)\n", "std_auc = np.std(aucs)\n", "plt.plot(mean_fpr, mean_tpr, color='b',\n", " label=r'Mean ROC (AUC = %0.2f $\\pm$ %0.2f)' % (mean_auc, std_auc),\n", " lw=2, alpha=.8)\n", "\n", "std_tpr = np.std(tprs, axis=0)\n", "tprs_upper = np.minimum(mean_tpr + std_tpr, 1)\n", "tprs_lower = np.maximum(mean_tpr - std_tpr, 0)\n", "plt.fill_between(mean_fpr, tprs_lower, tprs_upper, color='grey', alpha=.2,\n", " label=r'$\\pm$ 1 std. dev.')\n", "\n", "plt.xlim([-0.05, 1.05])\n", "plt.ylim([-0.05, 1.05])\n", "plt.xlabel('False Positive Rate')\n", "plt.ylabel('True Positive Rate')\n", "plt.title('ROC Curve for Citation Classification')\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "for train, test in cv.split(X, y):\n", " probas_ = pipeline.fit(X[train], y[train]).predict_proba(X[test])\n", " precision, recall, _ = precision_recall_curve(y[test], probas_[:,1])\n", "\n", " plt.step(recall, precision, alpha=0.2,\n", " where='post', label=f'average precision={average_precision_score(y[test], probas_[:,1])}')\n", "\n", " plt.xlabel('Recall')\n", " plt.ylabel('Precision')\n", " plt.ylim([0.0, 1.05])\n", " plt.xlim([0.0, 1.0])\n", "plt.title('Citation Precision-Recall curve'.format(\n", " average_precision_score(y[test], probas_[:,1])))\n", "plt.legend(loc=\"lower right\")\n", "plt.show()" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "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": "3.7.3" } }, "nbformat": 4, "nbformat_minor": 2 }