{ "cells": [ { "cell_type": "markdown", "id": "752115c7", "metadata": {}, "source": [ "## How to select predictors for the ESD model\n", "\n", "Here we domostrate how to use the predictor selection methods implemented in pyESD to select the set of predictor appropriate for a specific station and climate information predictand. \n", "\n", "We show the performace of the predictors selected by Recurssive, Sequential, and TreeBased methods and evaluate thier performace with a base model. \n" ] }, { "cell_type": "code", "execution_count": 2, "id": "1059f7b6", "metadata": {}, "outputs": [], "source": [ "# import all the models required\n", "import os \n", "import sys \n", "import pandas as pd \n", "import numpy as np \n", "from collections import OrderedDict\n", "import socket\n", "\n", "# modules related to pyESD\n", "\n", "from pyESD.Weatherstation import read_station_csv\n", "from pyESD.standardizer import MonthlyStandardizer\n", "from pyESD.ESD_utils import store_pickle, store_csv\n", "from pyESD.splitter import KFold\n", "from pyESD.ESD_utils import Dataset\n", "from pyESD.Weatherstation import read_weatherstationnames" ] }, { "cell_type": "markdown", "id": "d6874d74", "metadata": {}, "source": [ "### Data repositories\n", "To avoid the repetition of the paths to the predictors and predictand datasets, we have included them in a script that can easily be imported in all the notebooks\n", "1. The ERA5 datasets loaded with the Dataset module implemented in the ```pyESD.ESD_utils```\n", "2. The constructed time series of the predictors are stored in a pickle file to avoid loading them again on the next run. We set the predictordir to store the datasets\n", "3. We use the ```read_weatherstationnames``` to generated a list of all the weather station names that would be used to create the weather station objects\n", "4. We import all these variable as ```from read_data import * ``` which is this file: [read_data.py](./read_data.py)\n", "\n", "All the weather stations are stored in [data](./data)" ] }, { "cell_type": "markdown", "id": "eaa07ad6", "metadata": {}, "source": [ "### Predictors setting \n", "1. list of predictors to be loaded. Note that these names are the variables names stored in the netCDF files containing the predictors datasets\n", "2. set the date range that would be used to selected the predictors. Here it should be the same as the training and validation period\n" ] }, { "cell_type": "code", "execution_count": 22, "id": "4dd59b9b", "metadata": {}, "outputs": [], "source": [ "# potential predictor sets\n", "predictors = [\"t2m\", \"tp\",\"msl\", \"v10\", \"u10\", \n", " \"u250\", \"u850\", \"u500\",\"u700\", \"u1000\",\n", " \"v250\", \"v850\", \"v500\",\"v700\", \"v1000\",\n", " \"r250\", \"r850\", \"r500\",\"r700\", \"r1000\", \n", " \"z250\", \"z500\", \"z700\", \"z850\", \"z1000\", \n", " \"t250\", \"t850\", \"t500\",\"t700\", \"t1000\",\n", " \"dtd250\", \"dtd850\", \"dtd500\",\"dtd700\", \n", " \"dtd1000\"\n", " ]\n", "\n", "# training and validation period\n", "from1958to2010 = pd.date_range(start=\"1958-01-01\", end=\"2010-12-31\", freq=\"MS\")" ] }, { "cell_type": "markdown", "id": "b5e61c52", "metadata": {}, "source": [ "### Predictor selection experiment function\n", "Now we will create a function that would perform all the modeling routines so that we can easily loop through all the stations to perform the same task\n", "1. read the station data as object that would apply all the ESD routines\n", "2. set predictors with the list of predictors defined and the radius to construct the regional means\n", "3. standardize the data with any of the standardizers. Here we use the MonthlyStandardizer method\n", "4. defined the scoring metrics to be used for the validation\n", "5. set the model to be used for the ESD training (here we will use the Ridge CV model)\n", "6. fit the model, here we have to define the predictor selector method to be used for selecting the predictors\n", "7. get the selected predictors \n", "8. use the cross_validate_predict to get the cross-validation metrics of the model training \n", "9. store the selected predictors \n", "10. stored the validation metrics" ] }, { "cell_type": "code", "execution_count": 30, "id": "2ac96dbc", "metadata": {}, "outputs": [], "source": [ "\n", "def run_predictor_selection_example(variable, regressor, selector_method, cachedir, stationnames,\n", " station_datadir, predictors, predictordir, radius):\n", " \"\"\"\n", " Run an experiment using pyESD to perform predictor selection for a given variable.\n", "\n", " Args:\n", " variable (str): The target variable to predict, here Precipitation.\n", " regressor (str): The regression method to use, here we use the RidgeCV regression to test all the predictor selection\n", " methods.\n", " selector_method (str): The method for selecting predictors (\"Recursive\", \"TreeBased\", \"Sequential\").\n", " cachedir (str): Directory to store cached results, here all the files would be stored in the .\n", " stationnames (list): List of station names. it would be loaded from the read_data file\n", " station_datadir (str): Directory containing station data files: this is also set in the read the data file\n", " predictors (list): List of predictor variables.\n", " predictordir (str): Directory containing predictor data files.\n", " radius (float): Radius for selecting predictors: also defined in the read_data file\n", " \"\"\"\n", " num_of_stations = len(stationnames)\n", "\n", " # Loop through all stations\n", " for i in range(num_of_stations):\n", " stationname = stationnames[i]\n", " \n", " # set the exact path for the station data\n", " station_dir = os.path.join(station_datadir, stationname + \".csv\")\n", " \n", " # 1. create the station object using the read_station_csv and apply all the methods on the station object\n", " \n", " SO_instance = read_station_csv(filename=station_dir, varname=variable)\n", "\n", " # 2. Setting predictors (generate the predictors using the defined predictor names)\n", " SO_instance.set_predictors(variable, predictors, predictordir, radius)\n", "\n", " # 3. Setting standardizer\n", " SO_instance.set_standardizer(variable, standardizer=MonthlyStandardizer(detrending=False, scaling=False))\n", " \n", " # 4. define the scoring metrics\n", " scoring = [\"neg_root_mean_squared_error\", \"r2\", \"neg_mean_absolute_error\"]\n", " \n", " # 5. Setting model with cross-validation\n", " SO_instance.set_model(variable, method=regressor, scoring=scoring,\n", " cv=KFold(n_splits=10))\n", "\n", " # 6. Fitting model with predictor selector option\n", " if selector_method == \"Recursive\":\n", " SO_instance.fit(variable, from1958to2010, ERA5Data, fit_predictors=True, predictor_selector=True,\n", " selector_method=selector_method, selector_regressor=\"ARD\",\n", " cal_relative_importance=False)\n", " elif selector_method == \"TreeBased\":\n", " SO_instance.fit(variable, from1958to2010, ERA5Data, fit_predictors=True, predictor_selector=True,\n", " selector_method=selector_method, selector_regressor=\"RandomForest\")\n", " elif selector_method == \"Sequential\":\n", " SO_instance.fit(variable, from1958to2010, ERA5Data, fit_predictors=True, predictor_selector=True,\n", " selector_method=selector_method, selector_regressor=\"ARD\", num_predictors=10,\n", " selector_direction=\"forward\")\n", " else:\n", " raise ValueError(\"Define selector not recognized\")\n", "\n", " # 7. Extracting selected predictors\n", " selected_predictors = SO_instance.selected_names(variable)\n", "\n", " # 8. Training estimate for the same model\n", " \n", " score, ypred = SO_instance.cross_validate_and_predict(variable, from1958to2010, ERA5Data)\n", "\n", " # 9-10. Storing results using pickle\n", " store_pickle(stationname, \"selected_predictors_\" + selector_method, selected_predictors, cachedir)\n", " store_pickle(stationname, \"validation_score_\" + selector_method, score, cachedir)\n" ] }, { "cell_type": "markdown", "id": "6f29bb6f", "metadata": {}, "source": [ "### Perform the experimet \n", "We now use the above function to loop through the 5 stations to perform the experiment. \n", "\n", "#### import all the variables defined in the read_data module (e.g., paths to data, defined radius,)" ] }, { "cell_type": "code", "execution_count": 27, "id": "f237e0a4", "metadata": {}, "outputs": [], "source": [ "from read_data import radius, station_prec_datadir, stationnames_prec, ERA5Data, predictordir" ] }, { "cell_type": "markdown", "id": "e5592cd9", "metadata": {}, "source": [ "### Loop through the list of methods\n", "\n", "we now define the list of methods to test, and then loop through them to extract the predictors\n", "for all the methods and stations " ] }, { "cell_type": "code", "execution_count": 25, "id": "1e260f85", "metadata": {}, "outputs": [], "source": [ "selector_methods = [\"Recursive\", \"TreeBased\", \"Sequential\"]" ] }, { "cell_type": "code", "execution_count": 31, "id": "15b5d7a0", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "--------- running for method: Recursive --------\n", "Freiburg 48.0232 7.8343 236.0\n", "Regenerating predictor data for t2m using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for tp using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for msl using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "29 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'v10', 'u10', 'u250', 'u850', 'u500', 'u700', 'u1000',\n", " 'v250', 'v850', 'v500', 'v700', 'v1000', 'r250', 'r850', 'r500', 'r700',\n", " 'r1000', 'z1000', 't250', 't850', 't500', 't1000', 'dtd250', 'dtd850',\n", " 'dtd500', 'dtd700', 'dtd1000'],\n", " dtype='object')\n", "Konstanz 47.6952 9.1307 428.0\n", "Regenerating predictor data for t2m using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for tp using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for msl using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "35 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'msl', 'v10', 'u10', 'u250', 'u850', 'u500', 'u700',\n", " 'u1000', 'v250', 'v850', 'v500', 'v700', 'v1000', 'r250', 'r850',\n", " 'r500', 'r700', 'r1000', 'z250', 'z500', 'z700', 'z850', 'z1000',\n", " 't250', 't850', 't500', 't700', 't1000', 'dtd250', 'dtd850', 'dtd500',\n", " 'dtd700', 'dtd1000'],\n", " dtype='object')\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Mannheim 49.5063 8.5584 98.0\n", "Regenerating predictor data for t2m using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for tp using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for msl using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "6 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'u500', 'v250', 't500', 't700'], dtype='object')\n", "Nürnberg 49.503 11.0549 314.0\n", "Regenerating predictor data for t2m using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for tp using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for msl using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "33 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'msl', 'v10', 'u10', 'u250', 'u850', 'u500', 'u700',\n", " 'u1000', 'v250', 'v850', 'v500', 'v700', 'v1000', 'r250', 'r850',\n", " 'r500', 'r700', 'r1000', 'z250', 'z500', 'z700', 't250', 't850', 't500',\n", " 't700', 't1000', 'dtd250', 'dtd850', 'dtd500', 'dtd700', 'dtd1000'],\n", " dtype='object')\n", "Stuttgart (Schnarrenberg) 48.8281 9.2 314.0\n", "Regenerating predictor data for t2m using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Regenerating predictor data for tp using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for msl using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u10 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for u1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for v1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for r1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for z1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for t1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd250 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd850 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd500 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd700 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "Regenerating predictor data for dtd1000 using dataset ERA5 with loading patterns and params from ERA5 and ERA5\n", "31 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'msl', 'v10', 'u10', 'u250', 'u850', 'u500', 'u700',\n", " 'u1000', 'v850', 'v500', 'v700', 'v1000', 'r250', 'r850', 'r500',\n", " 'r700', 'r1000', 'z700', 'z1000', 't250', 't850', 't500', 't700',\n", " 't1000', 'dtd250', 'dtd850', 'dtd500', 'dtd700', 'dtd1000'],\n", " dtype='object')\n", "--------- running for method: TreeBased --------\n", "Freiburg 48.0232 7.8343 236.0\n", "1 : optimal number of predictors and selected variables are Index(['tp'], dtype='object')\n", "Konstanz 47.6952 9.1307 428.0\n", "2 : optimal number of predictors and selected variables are Index(['tp', 'u500'], dtype='object')\n", "Mannheim 49.5063 8.5584 98.0\n", "2 : optimal number of predictors and selected variables are Index(['tp', 'r1000'], dtype='object')\n", "Nürnberg 49.503 11.0549 314.0\n", "1 : optimal number of predictors and selected variables are Index(['tp'], dtype='object')\n", "Stuttgart (Schnarrenberg) 48.8281 9.2 314.0\n", "1 : optimal number of predictors and selected variables are Index(['tp'], dtype='object')\n", "--------- running for method: Sequential --------\n", "Freiburg 48.0232 7.8343 236.0\n", "10 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'u10', 'r250', 't250', 't700', 't1000', 'dtd250', 'dtd850',\n", " 'dtd1000'],\n", " dtype='object')\n", "Konstanz 47.6952 9.1307 428.0\n", "10 : optimal number of predictors and selected variables are Index(['tp', 'u10', 'u850', 'u700', 'u1000', 'v850', 'v500', 'v700', 'z250',\n", " 't250'],\n", " dtype='object')\n", "Mannheim 49.5063 8.5584 98.0\n", "10 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'msl', 'u10', 'u850', 'u500', 'r850', 't500', 't700',\n", " 'dtd1000'],\n", " dtype='object')\n", "Nürnberg 49.503 11.0549 314.0\n", "10 : optimal number of predictors and selected variables are Index(['t2m', 'tp', 'u10', 'u250', 'u700', 'v850', 'v500', 'r850', 'z1000',\n", " 'dtd700'],\n", " dtype='object')\n", "Stuttgart (Schnarrenberg) 48.8281 9.2 314.0\n", "10 : optimal number of predictors and selected variables are Index(['tp', 'v10', 'u10', 'v250', 'v1000', 'r1000', 'z250', 't250', 'dtd700',\n", " 'dtd1000'],\n", " dtype='object')\n" ] } ], "source": [ "for selector in selector_methods:\n", " print (\"--------- running for method: \", selector, \"--------\")\n", " run_predictor_selection_example(variable=\"Precipitation\", regressor=\"RidgeCV\", selector_method=selector,\n", " cachedir=cachedir_prec, stationnames=stationnames_prec,\n", " station_datadir=station_prec_datadir, predictors=predictors, \n", " predictordir=predictordir, radius=radius)" ] }, { "cell_type": "markdown", "id": "be60fbfd", "metadata": {}, "source": [ "### Checking the results\n", "Oops! These experiments took a while to run, now its time to check if the results make sense and whether we can decide on which\n", "predictor selection method to use for our modelling routines in the next experiments. \n", "\n", "All the results data are stored in the [predictor selection results folder](./predictor_selection_results)\n", "\n", "Now we will use the functions in the ```pyESD.plot_utils``` and ```pyESD.plot``` to visualize the results\n" ] }, { "cell_type": "code", "execution_count": 33, "id": "3d0cc92a", "metadata": {}, "outputs": [], "source": [ "#path to the results\n", "path_to_results = \"C:/Users/dboateng/Desktop/Python_scripts/ESD_Package/examples/tutorials/predictor_selection_results\"\n" ] }, { "cell_type": "markdown", "id": "c04ebdaa", "metadata": {}, "source": [ "#### The RidgeCV model performance using the predictor selection methods\n", "1. We plot the R² of all the stations using the 3 methods tested\n", "2. We plot the RMSE" ] }, { "cell_type": "code", "execution_count": 39, "id": "4199e1d3", "metadata": {}, "outputs": [], "source": [ "from pyESD.plot import barplot" ] }, { "cell_type": "code", "execution_count": 43, "id": "920e324b", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "barplot(methods=selector_methods, stationnames=stationnames_prec , path_to_data=path_to_results, \n", " xlabel=\"Stations\", ylabel=\"CV R²\", varname= \"test_r2\", varname_std =\"test_r2_std\",\n", " filename=\"validation_score_\", legend=True,)" ] }, { "cell_type": "markdown", "id": "d223746e", "metadata": {}, "source": [ "#### Discussion of the results \n", "Fantastic! Here, we can see that using a simple model (RidgeCV) with the selected predictors from the differnt methods show the differences in the individual stations. Even though the difference are not that significant, we can see that Treebased method performs worse as compared to the others for most of the station. The difference between the Recursive and Sequential is very small, therefore we can confidently select one" ] }, { "cell_type": "code", "execution_count": 35, "id": "e7372d5e", "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": { "needs_background": "light" }, "output_type": "display_data" } ], "source": [ "barplot(methods=selector_methods, stationnames=stationnames_prec , path_to_data=path_to_results, \n", " xlabel=\"Stations\", ylabel=\"CV RMSE\", varname= \"test_rmse\", varname_std =\"test_rmse_std\",\n", " filename=\"validation_score_\", )" ] }, { "cell_type": "markdown", "id": "a85cfd1e", "metadata": {}, "source": [ "In terms of the RMSE, the results shows consistent pattern. Therefore, we will use Recursive for the subsequent experiments\n" ] }, { "cell_type": "markdown", "id": "3978c28a", "metadata": {}, "source": [ "## We can also count the predictors selected \n", "Here we use the ```count_predictors``` function in the ```pyESD.plot_utils```" ] }, { "cell_type": "code", "execution_count": 45, "id": "818c3abc", "metadata": {}, "outputs": [], "source": [ "from pyESD.plot_utils import count_predictors" ] }, { "cell_type": "code", "execution_count": 46, "id": "627e3835", "metadata": {}, "outputs": [], "source": [ "df = count_predictors(selector_methods, stationnames_prec, path_to_results,\n", " \"selected_predictors_\", predictors)" ] }, { "cell_type": "code", "execution_count": 47, "id": "d8129f43", "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", "
t2mtpmslv10u10u250u850u500u700u1000...t250t850t500t700t1000dtd250dtd850dtd500dtd700dtd1000
Recursive5.05.03.04.04.04.04.05.04.04.0...4.04.05.04.04.04.04.04.04.04.0
TreeBasedNaN5.0NaNNaNNaNNaNNaN1.0NaNNaN...NaNNaNNaNNaNNaNNaNNaNNaNNaNNaN
Sequential3.05.01.01.05.01.02.01.02.01.0...3.0NaN1.02.01.01.01.0NaN2.03.0
\n", "

3 rows × 35 columns

\n", "
" ], "text/plain": [ " t2m tp msl v10 u10 u250 u850 u500 u700 u1000 ... t250 \\\n", "Recursive 5.0 5.0 3.0 4.0 4.0 4.0 4.0 5.0 4.0 4.0 ... 4.0 \n", "TreeBased NaN 5.0 NaN NaN NaN NaN NaN 1.0 NaN NaN ... NaN \n", "Sequential 3.0 5.0 1.0 1.0 5.0 1.0 2.0 1.0 2.0 1.0 ... 3.0 \n", "\n", " t850 t500 t700 t1000 dtd250 dtd850 dtd500 dtd700 dtd1000 \n", "Recursive 4.0 5.0 4.0 4.0 4.0 4.0 4.0 4.0 4.0 \n", "TreeBased NaN NaN NaN NaN NaN NaN NaN NaN NaN \n", "Sequential NaN 1.0 2.0 1.0 1.0 1.0 NaN 2.0 3.0 \n", "\n", "[3 rows x 35 columns]" ] }, "execution_count": 47, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df" ] }, { "cell_type": "markdown", "id": "7a8eaf72", "metadata": {}, "source": [ "This would provide the frequency of the variables that was selected in each of the methods. Eg. First column: Temperature was selected 5x for Recursive, 0 for TreeBased and 3 for Sequential" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "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": "3.9.7" } }, "nbformat": 4, "nbformat_minor": 5 }