From fceb2c729fc443b0ce6cce9a9d4abc0cbcdc665d Mon Sep 17 00:00:00 2001 From: zachrewolinski Date: Wed, 15 May 2024 10:39:15 -0700 Subject: [PATCH] lets see if this goes through --- feature_importance/feature_ranking.sh | 2 +- .../diabetes-regression/linear-model/dgp.py | 6 +- .../linear-model/models.py | 2 +- .../ranking_importance_local_sims.ipynb | 2062 ++++++++++------- .../ranking_importance_local_sims.py | 9 +- .../scripts/competing_methods_local.py | 40 +- 6 files changed, 1260 insertions(+), 861 deletions(-) diff --git a/feature_importance/feature_ranking.sh b/feature_importance/feature_ranking.sh index da22e1e..85dc891 100644 --- a/feature_importance/feature_ranking.sh +++ b/feature_importance/feature_ranking.sh @@ -3,7 +3,7 @@ #SBATCH --mail-type=ALL source activate mdi -command="ranking_importance_local_sims.py --nreps 1 --config mdi_local.real_x_sim_y.diabetes-regression.linear-model --split_seed ${1} --ignore_cache --create_rmd --result_name diabetes-reg-linear" +command="ranking_importance_local_sims.py --nreps 1 --config mdi_local.real_x_sim_y.diabetes-regression.linear-model --split_seed 1 --ignore_cache --create_rmd --result_name diabetes-reg-linear" # Execute the command python $command \ No newline at end of file diff --git a/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/dgp.py b/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/dgp.py index e57ac84..c21746a 100644 --- a/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/dgp.py +++ b/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/dgp.py @@ -7,7 +7,8 @@ X_PARAMS_DICT = { "source": "imodels", "data_name": "diabetes_regr", - "sample_row_n": None + "sample_row_n": 400 + #"sample_row_n": None } Y_DGP = linear_model @@ -25,5 +26,4 @@ # "300": 300, "400": 400}} VARY_PARAM_NAME = ["heritability"] -VARY_PARAM_VALS = {"heritability": {"0.1": 0.1, "0.2": 0.2, - "0.4": 0.4, "0.8": 0.8}} +VARY_PARAM_VALS = {"heritability": {"0.4": 0.4, "0.8": 0.8}} diff --git a/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/models.py b/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/models.py index 5fdb58e..8b13107 100644 --- a/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/models.py +++ b/feature_importance/fi_config/mdi_local/real_x_sim_y/diabetes-regression/linear-model/models.py @@ -7,7 +7,7 @@ ESTIMATORS = [ [ModelConfig('RF', RandomForestRegressor, model_type='tree', - other_params={'n_estimators': 100, 'min_samples_leaf': 1, 'max_features': 0.33, 'random_state': 42})] + other_params={'n_estimators': 100, 'min_samples_leaf': 5, 'max_features': 0.33, 'random_state': 42})] ] FI_ESTIMATORS = [ diff --git a/feature_importance/ranking_importance_local_sims.ipynb b/feature_importance/ranking_importance_local_sims.ipynb index 282ff64..937d02d 100644 --- a/feature_importance/ranking_importance_local_sims.ipynb +++ b/feature_importance/ranking_importance_local_sims.ipynb @@ -2,7 +2,7 @@ "cells": [ { "cell_type": "code", - "execution_count": 3, + "execution_count": 1, "metadata": {}, "outputs": [], "source": [ @@ -15,7 +15,7 @@ }, { "cell_type": "code", - "execution_count": 4, + "execution_count": 2, "metadata": {}, "outputs": [], "source": [ @@ -34,7 +34,7 @@ }, { "cell_type": "code", - "execution_count": 5, + "execution_count": 3, "metadata": {}, "outputs": [ { @@ -193,26 +193,437 @@ " test_F1\n", " split_seed\n", " \n", - " \n", - " \n", + " \n", + " \n", + " \n", + " 0\n", + " 0\n", + " 0.4\n", + " 0.4\n", + " 100\n", + " 5\n", + " 0.33\n", + " 42\n", + " NaN\n", + " NaN\n", + " RF\n", + " Kernel_SHAP_RF_plus\n", + " 296\n", + " 146\n", + " 10\n", + " 1\n", + " 2.749968\n", + " 274\n", + " 155\n", + " 84\n", + " 82\n", + " 261\n", + " 9\n", + " 42\n", + " 277\n", + " 282\n", + " 92\n", + " 148\n", + " 211\n", + " 60\n", + " 218\n", + " 262\n", + " 46\n", + " 45\n", + " 236\n", + " 228\n", + " 132\n", + " 143\n", + " 167\n", + " 152\n", + " 93\n", + " 113\n", + " 5\n", + " 238\n", + " 251\n", + " 170\n", + " 186\n", + " 193\n", + " 33\n", + " 222\n", + " 216\n", + " 197\n", + " 73\n", + " 182\n", + " 119\n", + " 285\n", + " 202\n", + " 204\n", + " 179\n", + " 177\n", + " 111\n", + " 59\n", + " 226\n", + " 25\n", + " 77\n", + " 6\n", + " 175\n", + " 164\n", + " 140\n", + " 30\n", + " 22\n", + " 245\n", + " 24\n", + " 56\n", + " 144\n", + " 124\n", + " 97\n", + " 63\n", + " 17\n", + " 215\n", + " 219\n", + " 183\n", + " 114\n", + " 76\n", + " 284\n", + " 66\n", + " 178\n", + " 154\n", + " 75\n", + " 19\n", + " 108\n", + " 69\n", + " 30\n", + " 39\n", + " 2\n", + " 124\n", + " 10\n", + " 68\n", + " 51\n", + " 71\n", + " 77\n", + " 102\n", + " 80\n", + " 76\n", + " 142\n", + " 127\n", + " 95\n", + " 70\n", + " 93\n", + " 67\n", + " 0\n", + " 105\n", + " 82\n", + " 136\n", + " 40\n", + " 54\n", + " 28\n", + " 74\n", + " 119\n", + " 18\n", + " 9\n", + " 58\n", + " 99\n", + " 73\n", + " 97\n", + " 128\n", + " 122\n", + " 157.653250\n", + " 0.884865\n", + " 0.907278\n", + " 0.0\n", + " 0.930000\n", + " 0.941834\n", + " 0.0\n", + " 1\n", + " \n", + " \n", + " 1\n", + " 0\n", + " 0.4\n", + " 0.4\n", + " 100\n", + " 5\n", + " 0.33\n", + " 42\n", + " NaN\n", + " NaN\n", + " RF\n", + " LFI_evaluate_on_all_RF_plus\n", + " 296\n", + " 146\n", + " 10\n", + " 1\n", + " 66.605561\n", + " 274\n", + " 155\n", + " 84\n", + " 82\n", + " 261\n", + " 9\n", + " 42\n", + " 277\n", + " 282\n", + " 92\n", + " 148\n", + " 211\n", + " 60\n", + " 218\n", + " 262\n", + " 46\n", + " 45\n", + " 236\n", + " 228\n", + " 132\n", + " 143\n", + " 167\n", + " 152\n", + " 93\n", + " 113\n", + " 5\n", + " 238\n", + " 251\n", + " 170\n", + " 186\n", + " 193\n", + " 33\n", + " 222\n", + " 216\n", + " 197\n", + " 73\n", + " 182\n", + " 119\n", + " 285\n", + " 202\n", + " 204\n", + " 179\n", + " 177\n", + " 111\n", + " 59\n", + " 226\n", + " 25\n", + " 77\n", + " 6\n", + " 175\n", + " 164\n", + " 140\n", + " 30\n", + " 22\n", + " 245\n", + " 24\n", + " 56\n", + " 144\n", + " 124\n", + " 97\n", + " 63\n", + " 17\n", + " 215\n", + " 219\n", + " 183\n", + " 114\n", + " 76\n", + " 284\n", + " 66\n", + " 178\n", + " 154\n", + " 75\n", + " 19\n", + " 108\n", + " 69\n", + " 30\n", + " 39\n", + " 2\n", + " 124\n", + " 10\n", + " 68\n", + " 51\n", + " 71\n", + " 77\n", + " 102\n", + " 80\n", + " 76\n", + " 142\n", + " 127\n", + " 95\n", + " 70\n", + " 93\n", + " 67\n", + " 0\n", + " 105\n", + " 82\n", + " 136\n", + " 40\n", + " 54\n", + " 28\n", + " 74\n", + " 119\n", + " 18\n", + " 9\n", + " 58\n", + " 99\n", + " 73\n", + " 97\n", + " 128\n", + " 122\n", + " 2.366904\n", + " 0.590270\n", + " 0.731816\n", + " 0.0\n", + " 0.574444\n", + " 0.724134\n", + " 0.0\n", + " 1\n", + " \n", + " \n", + " 2\n", + " 0\n", + " 0.4\n", + " 0.4\n", + " 100\n", + " 5\n", + " 0.33\n", + " 42\n", + " NaN\n", + " NaN\n", + " RF\n", + " LFI_evaluate_on_oob_RF_plus\n", + " 296\n", + " 146\n", + " 10\n", + " 1\n", + " 2.366904\n", + " 274\n", + " 155\n", + " 84\n", + " 82\n", + " 261\n", + " 9\n", + " 42\n", + " 277\n", + " 282\n", + " 92\n", + " 148\n", + " 211\n", + " 60\n", + " 218\n", + " 262\n", + " 46\n", + " 45\n", + " 236\n", + " 228\n", + " 132\n", + " 143\n", + " 167\n", + " 152\n", + " 93\n", + " 113\n", + " 5\n", + " 238\n", + " 251\n", + " 170\n", + " 186\n", + " 193\n", + " 33\n", + " 222\n", + " 216\n", + " 197\n", + " 73\n", + " 182\n", + " 119\n", + " 285\n", + " 202\n", + " 204\n", + " 179\n", + " 177\n", + " 111\n", + " 59\n", + " 226\n", + " 25\n", + " 77\n", + " 6\n", + " 175\n", + " 164\n", + " 140\n", + " 30\n", + " 22\n", + " 245\n", + " 24\n", + " 56\n", + " 144\n", + " 124\n", + " 97\n", + " 63\n", + " 17\n", + " 215\n", + " 219\n", + " 183\n", + " 114\n", + " 76\n", + " 284\n", + " 66\n", + " 178\n", + " 154\n", + " 75\n", + " 19\n", + " 108\n", + " 69\n", + " 30\n", + " 39\n", + " 2\n", + " 124\n", + " 10\n", + " 68\n", + " 51\n", + " 71\n", + " 77\n", + " 102\n", + " 80\n", + " 76\n", + " 142\n", + " 127\n", + " 95\n", + " 70\n", + " 93\n", + " 67\n", + " 0\n", + " 105\n", + " 82\n", + " 136\n", + " 40\n", + " 54\n", + " 28\n", + " 74\n", + " 119\n", + " 18\n", + " 9\n", + " 58\n", + " 99\n", + " 73\n", + " 97\n", + " 128\n", + " 122\n", + " 2.749968\n", + " 0.574054\n", + " 0.722772\n", + " 0.0\n", + " 0.574444\n", + " 0.724134\n", + " 0.0\n", + " 1\n", + " \n", " \n", - " 0\n", + " 3\n", " 0\n", - " 0.1\n", - " 0.1\n", + " 0.4\n", + " 0.4\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", - " NaN\n", + " oob\n", " RF\n", - " Kernel_SHAP_RF_plus\n", + " LFI_fit_on_OOB_RF\n", " 296\n", " 146\n", " 10\n", " 1\n", - " 12.777327\n", + " 4.271746\n", " 274\n", " 155\n", " 84\n", @@ -323,33 +734,170 @@ " 97\n", " 128\n", " 122\n", - " 1622.727949\n", - " 0.577838\n", - " 0.635698\n", - " 0.000000\n", - " 0.646667\n", - " 0.677123\n", - " 0.000000\n", + " 66.605561\n", + " 0.581081\n", + " 0.729977\n", + " 0.0\n", + " 0.597778\n", + " 0.743177\n", + " 0.0\n", " 1\n", " \n", " \n", - " 1\n", + " 4\n", " 0\n", - " 0.1\n", - " 0.1\n", + " 0.4\n", + " 0.4\n", " 100\n", + " 5\n", + " 0.33\n", + " 42\n", + " False\n", + " inbag\n", + " RF\n", + " LFI_fit_on_inbag_RF\n", + " 296\n", + " 146\n", + " 10\n", + " 1\n", + " 0.414457\n", + " 274\n", + " 155\n", + " 84\n", + " 82\n", + " 261\n", + " 9\n", + " 42\n", + " 277\n", + " 282\n", + " 92\n", + " 148\n", + " 211\n", + " 60\n", + " 218\n", + " 262\n", + " 46\n", + " 45\n", + " 236\n", + " 228\n", + " 132\n", + " 143\n", + " 167\n", + " 152\n", + " 93\n", + " 113\n", + " 5\n", + " 238\n", + " 251\n", + " 170\n", + " 186\n", + " 193\n", + " 33\n", + " 222\n", + " 216\n", + " 197\n", + " 73\n", + " 182\n", + " 119\n", + " 285\n", + " 202\n", + " 204\n", + " 179\n", + " 177\n", + " 111\n", + " 59\n", + " 226\n", + " 25\n", + " 77\n", + " 6\n", + " 175\n", + " 164\n", + " 140\n", + " 30\n", + " 22\n", + " 245\n", + " 24\n", + " 56\n", + " 144\n", + " 124\n", + " 97\n", + " 63\n", + " 17\n", + " 215\n", + " 219\n", + " 183\n", + " 114\n", + " 76\n", + " 284\n", + " 66\n", + " 178\n", + " 154\n", + " 75\n", + " 19\n", + " 108\n", + " 69\n", + " 30\n", + " 39\n", + " 2\n", + " 124\n", + " 10\n", + " 68\n", + " 51\n", + " 71\n", + " 77\n", + " 102\n", + " 80\n", + " 76\n", + " 142\n", + " 127\n", + " 95\n", + " 70\n", + " 93\n", + " 67\n", + " 0\n", + " 105\n", + " 82\n", + " 136\n", + " 40\n", + " 54\n", + " 28\n", + " 74\n", + " 119\n", + " 18\n", + " 9\n", + " 58\n", + " 99\n", + " 73\n", + " 97\n", + " 128\n", + " 122\n", + " 4.271746\n", + " 0.537297\n", + " 0.667003\n", + " 0.0\n", + " 0.524444\n", + " 0.662643\n", + " 0.0\n", " 1\n", + " \n", + " \n", + " 5\n", + " 0\n", + " 0.4\n", + " 0.4\n", + " 100\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", " NaN\n", " RF\n", - " LFI_evaluate_on_all_RF_plus\n", + " LIME_RF_plus\n", " 296\n", " 146\n", " 10\n", " 1\n", - " 0.000003\n", + " 157.653250\n", " 274\n", " 155\n", " 84\n", @@ -460,33 +1008,33 @@ " 97\n", " 128\n", " 122\n", - " 11.966596\n", - " 0.495676\n", - " 0.612882\n", - " 0.130066\n", - " 0.451111\n", - " 0.581673\n", - " 0.013889\n", + " 170.402314\n", + " 0.876216\n", + " 0.905993\n", + " 0.0\n", + " 0.877778\n", + " 0.909645\n", + " 0.0\n", " 1\n", " \n", " \n", - " 2\n", + " 6\n", " 0\n", - " 0.1\n", - " 0.1\n", + " 0.4\n", + " 0.4\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", " NaN\n", " RF\n", - " LFI_evaluate_on_oob_RF_plus\n", + " TreeSHAP_RF\n", " 296\n", " 146\n", " 10\n", " 1\n", - " 11.966596\n", + " 58.843677\n", " 274\n", " 155\n", " 84\n", @@ -597,33 +1145,33 @@ " 97\n", " 128\n", " 122\n", - " 12.777327\n", - " 0.474595\n", - " 0.598768\n", - " 0.149544\n", - " 0.451111\n", - " 0.581673\n", - " 0.013889\n", + " 0.414457\n", + " 0.805946\n", + " 0.821571\n", + " 0.0\n", + " 0.846667\n", + " 0.862467\n", + " 0.0\n", " 1\n", " \n", " \n", - " 3\n", + " 7\n", " 0\n", - " 0.1\n", - " 0.1\n", + " 0.8\n", + " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", - " oob\n", + " NaN\n", " RF\n", - " LFI_fit_on_OOB_RF\n", + " Kernel_SHAP_RF_plus\n", " 296\n", " 146\n", " 10\n", " 1\n", - " 23.028107\n", + " 2.728971\n", " 274\n", " 155\n", " 84\n", @@ -734,33 +1282,33 @@ " 97\n", " 128\n", " 122\n", - " 0.000003\n", - " 0.521622\n", - " 0.638607\n", - " 0.101366\n", - " 0.476667\n", - " 0.588973\n", - " 0.029101\n", + " 158.937756\n", + " 0.954595\n", + " 0.967038\n", + " 0.0\n", + " 0.983333\n", + " 0.988194\n", + " 0.0\n", " 1\n", " \n", " \n", - " 4\n", + " 8\n", " 0\n", - " 0.1\n", - " 0.1\n", + " 0.8\n", + " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", - " False\n", - " inbag\n", + " NaN\n", + " NaN\n", " RF\n", - " LFI_fit_on_inbag_RF\n", + " LFI_evaluate_on_all_RF_plus\n", " 296\n", " 146\n", " 10\n", " 1\n", - " 4.546044\n", + " 63.704902\n", " 274\n", " 155\n", " 84\n", @@ -871,159 +1419,22 @@ " 97\n", " 128\n", " 122\n", - " 23.028107\n", - " 0.521622\n", - " 0.638607\n", - " 0.101366\n", - " 0.476667\n", - " 0.588973\n", - " 0.029101\n", + " 2.351976\n", + " 0.643243\n", + " 0.788523\n", + " 0.0\n", + " 0.594444\n", + " 0.759802\n", + " 0.0\n", " 1\n", " \n", " \n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " ...\n", - " \n", - " \n", - " 275\n", + " 9\n", " 0\n", " 0.8\n", " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", @@ -1033,8 +1444,8 @@ " 296\n", " 146\n", " 10\n", - " 3\n", - " 10.015577\n", + " 1\n", + " 2.351976\n", " 274\n", " 155\n", " 84\n", @@ -1145,22 +1556,22 @@ " 97\n", " 128\n", " 122\n", - " 10.731668\n", - " 0.327568\n", - " 0.573769\n", - " 0.000000\n", - " 0.272222\n", - " 0.543042\n", - " 0.000000\n", - " 3\n", + " 2.728971\n", + " 0.637838\n", + " 0.785668\n", + " 0.0\n", + " 0.594444\n", + " 0.759802\n", + " 0.0\n", + " 1\n", " \n", " \n", - " 276\n", + " 10\n", " 0\n", " 0.8\n", " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", @@ -1170,8 +1581,8 @@ " 296\n", " 146\n", " 10\n", - " 3\n", - " 19.305907\n", + " 1\n", + " 4.068172\n", " 274\n", " 155\n", " 84\n", @@ -1282,22 +1693,22 @@ " 97\n", " 128\n", " 122\n", - " 0.000002\n", - " 0.433514\n", - " 0.614721\n", - " 0.000000\n", - " 0.410000\n", - " 0.596803\n", - " 0.000000\n", - " 3\n", + " 63.704902\n", + " 0.650811\n", + " 0.798041\n", + " 0.0\n", + " 0.598889\n", + " 0.764195\n", + " 0.0\n", + " 1\n", " \n", " \n", - " 277\n", + " 11\n", " 0\n", " 0.8\n", " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " False\n", @@ -1307,8 +1718,8 @@ " 296\n", " 146\n", " 10\n", - " 3\n", - " 4.426196\n", + " 1\n", + " 0.400583\n", " 274\n", " 155\n", " 84\n", @@ -1419,22 +1830,22 @@ " 97\n", " 128\n", " 122\n", - " 19.305907\n", - " 0.433514\n", - " 0.614721\n", - " 0.000000\n", - " 0.410000\n", - " 0.596803\n", - " 0.000000\n", - " 3\n", + " 4.068172\n", + " 0.550811\n", + " 0.684134\n", + " 0.0\n", + " 0.521111\n", + " 0.667200\n", + " 0.0\n", + " 1\n", " \n", " \n", - " 278\n", + " 12\n", " 0\n", " 0.8\n", " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", @@ -1444,8 +1855,8 @@ " 296\n", " 146\n", " 10\n", - " 3\n", - " 1339.517789\n", + " 1\n", + " 158.937756\n", " 274\n", " 155\n", " 84\n", @@ -1556,22 +1967,22 @@ " 97\n", " 128\n", " 122\n", - " 1775.899545\n", - " 0.946486\n", - " 0.957699\n", - " 0.000000\n", - " 0.952222\n", - " 0.962438\n", - " 0.000000\n", - " 3\n", + " 172.658590\n", + " 0.984865\n", + " 0.986969\n", + " 0.0\n", + " 0.983333\n", + " 0.986455\n", + " 0.0\n", + " 1\n", " \n", " \n", - " 279\n", + " 13\n", " 0\n", " 0.8\n", " 0.8\n", " 100\n", - " 1\n", + " 5\n", " 0.33\n", " 42\n", " NaN\n", @@ -1581,8 +1992,8 @@ " 296\n", " 146\n", " 10\n", - " 3\n", - " 387.387989\n", + " 1\n", + " 56.777046\n", " 274\n", " 155\n", " 84\n", @@ -1693,454 +2104,534 @@ " 97\n", " 128\n", " 122\n", - " 4.426196\n", - " 0.848108\n", - " 0.888083\n", - " 0.000000\n", - " 0.828889\n", - " 0.872943\n", - " 0.000000\n", - " 3\n", + " 0.400583\n", + " 0.793514\n", + " 0.831404\n", + " 0.0\n", + " 0.831111\n", + " 0.862800\n", + " 0.0\n", + " 1\n", " \n", " \n", "\n", - "

280 rows × 134 columns

\n", "" ], "text/plain": [ - " rep heritability heritability_name n_estimators min_samples_leaf \\\n", - "0 0 0.1 0.1 100 1 \n", - "1 0 0.1 0.1 100 1 \n", - "2 0 0.1 0.1 100 1 \n", - "3 0 0.1 0.1 100 1 \n", - "4 0 0.1 0.1 100 1 \n", - ".. ... ... ... ... ... \n", - "275 0 0.8 0.8 100 1 \n", - "276 0 0.8 0.8 100 1 \n", - "277 0 0.8 0.8 100 1 \n", - "278 0 0.8 0.8 100 1 \n", - "279 0 0.8 0.8 100 1 \n", + " rep heritability heritability_name n_estimators min_samples_leaf \\\n", + "0 0 0.4 0.4 100 5 \n", + "1 0 0.4 0.4 100 5 \n", + "2 0 0.4 0.4 100 5 \n", + "3 0 0.4 0.4 100 5 \n", + "4 0 0.4 0.4 100 5 \n", + "5 0 0.4 0.4 100 5 \n", + "6 0 0.4 0.4 100 5 \n", + "7 0 0.8 0.8 100 5 \n", + "8 0 0.8 0.8 100 5 \n", + "9 0 0.8 0.8 100 5 \n", + "10 0 0.8 0.8 100 5 \n", + "11 0 0.8 0.8 100 5 \n", + "12 0 0.8 0.8 100 5 \n", + "13 0 0.8 0.8 100 5 \n", "\n", - " max_features random_state include_raw fit_on model \\\n", - "0 0.33 42 NaN NaN RF \n", - "1 0.33 42 NaN NaN RF \n", - "2 0.33 42 NaN NaN RF \n", - "3 0.33 42 NaN oob RF \n", - "4 0.33 42 False inbag RF \n", - ".. ... ... ... ... ... \n", - "275 0.33 42 NaN NaN RF \n", - "276 0.33 42 NaN oob RF \n", - "277 0.33 42 False inbag RF \n", - "278 0.33 42 NaN NaN RF \n", - "279 0.33 42 NaN NaN RF \n", + " max_features random_state include_raw fit_on model \\\n", + "0 0.33 42 NaN NaN RF \n", + "1 0.33 42 NaN NaN RF \n", + "2 0.33 42 NaN NaN RF \n", + "3 0.33 42 NaN oob RF \n", + "4 0.33 42 False inbag RF \n", + "5 0.33 42 NaN NaN RF \n", + "6 0.33 42 NaN NaN RF \n", + "7 0.33 42 NaN NaN RF \n", + "8 0.33 42 NaN NaN RF \n", + "9 0.33 42 NaN NaN RF \n", + "10 0.33 42 NaN oob RF \n", + "11 0.33 42 False inbag RF \n", + "12 0.33 42 NaN NaN RF \n", + "13 0.33 42 NaN NaN RF \n", "\n", - " fi train_size test_size num_features \\\n", - "0 Kernel_SHAP_RF_plus 296 146 10 \n", - "1 LFI_evaluate_on_all_RF_plus 296 146 10 \n", - "2 LFI_evaluate_on_oob_RF_plus 296 146 10 \n", - "3 LFI_fit_on_OOB_RF 296 146 10 \n", - "4 LFI_fit_on_inbag_RF 296 146 10 \n", - ".. ... ... ... ... \n", - "275 LFI_evaluate_on_oob_RF_plus 296 146 10 \n", - "276 LFI_fit_on_OOB_RF 296 146 10 \n", - "277 LFI_fit_on_inbag_RF 296 146 10 \n", - "278 LIME_RF_plus 296 146 10 \n", - "279 TreeSHAP_RF 296 146 10 \n", + " fi train_size test_size num_features \\\n", + "0 Kernel_SHAP_RF_plus 296 146 10 \n", + "1 LFI_evaluate_on_all_RF_plus 296 146 10 \n", + "2 LFI_evaluate_on_oob_RF_plus 296 146 10 \n", + "3 LFI_fit_on_OOB_RF 296 146 10 \n", + "4 LFI_fit_on_inbag_RF 296 146 10 \n", + "5 LIME_RF_plus 296 146 10 \n", + "6 TreeSHAP_RF 296 146 10 \n", + "7 Kernel_SHAP_RF_plus 296 146 10 \n", + "8 LFI_evaluate_on_all_RF_plus 296 146 10 \n", + "9 LFI_evaluate_on_oob_RF_plus 296 146 10 \n", + "10 LFI_fit_on_OOB_RF 296 146 10 \n", + "11 LFI_fit_on_inbag_RF 296 146 10 \n", + "12 LIME_RF_plus 296 146 10 \n", + "13 TreeSHAP_RF 296 146 10 \n", "\n", - " data_split_seed rf_plus_fit_time sample_train_0 sample_train_1 \\\n", - "0 1 12.777327 274 155 \n", - "1 1 0.000003 274 155 \n", - "2 1 11.966596 274 155 \n", - "3 1 23.028107 274 155 \n", - "4 1 4.546044 274 155 \n", - ".. ... ... ... ... \n", - "275 3 10.015577 274 155 \n", - "276 3 19.305907 274 155 \n", - "277 3 4.426196 274 155 \n", - "278 3 1339.517789 274 155 \n", - "279 3 387.387989 274 155 \n", + " data_split_seed rf_plus_fit_time sample_train_0 sample_train_1 \\\n", + "0 1 2.749968 274 155 \n", + "1 1 66.605561 274 155 \n", + "2 1 2.366904 274 155 \n", + "3 1 4.271746 274 155 \n", + "4 1 0.414457 274 155 \n", + "5 1 157.653250 274 155 \n", + "6 1 58.843677 274 155 \n", + "7 1 2.728971 274 155 \n", + "8 1 63.704902 274 155 \n", + "9 1 2.351976 274 155 \n", + "10 1 4.068172 274 155 \n", + "11 1 0.400583 274 155 \n", + "12 1 158.937756 274 155 \n", + "13 1 56.777046 274 155 \n", "\n", - " sample_train_2 sample_train_3 sample_train_4 sample_train_5 \\\n", - "0 84 82 261 9 \n", - "1 84 82 261 9 \n", - "2 84 82 261 9 \n", - "3 84 82 261 9 \n", - "4 84 82 261 9 \n", - ".. ... ... ... ... \n", - "275 84 82 261 9 \n", - "276 84 82 261 9 \n", - "277 84 82 261 9 \n", - "278 84 82 261 9 \n", - "279 84 82 261 9 \n", + " sample_train_2 sample_train_3 sample_train_4 sample_train_5 \\\n", + "0 84 82 261 9 \n", + "1 84 82 261 9 \n", + "2 84 82 261 9 \n", + "3 84 82 261 9 \n", + "4 84 82 261 9 \n", + "5 84 82 261 9 \n", + "6 84 82 261 9 \n", + "7 84 82 261 9 \n", + "8 84 82 261 9 \n", + "9 84 82 261 9 \n", + "10 84 82 261 9 \n", + "11 84 82 261 9 \n", + "12 84 82 261 9 \n", + "13 84 82 261 9 \n", "\n", - " sample_train_6 sample_train_7 sample_train_8 sample_train_9 \\\n", - "0 42 277 282 92 \n", - "1 42 277 282 92 \n", - "2 42 277 282 92 \n", - "3 42 277 282 92 \n", - "4 42 277 282 92 \n", - ".. ... ... ... ... \n", - "275 42 277 282 92 \n", - "276 42 277 282 92 \n", - "277 42 277 282 92 \n", - "278 42 277 282 92 \n", - "279 42 277 282 92 \n", + " sample_train_6 sample_train_7 sample_train_8 sample_train_9 \\\n", + "0 42 277 282 92 \n", + "1 42 277 282 92 \n", + "2 42 277 282 92 \n", + "3 42 277 282 92 \n", + "4 42 277 282 92 \n", + "5 42 277 282 92 \n", + "6 42 277 282 92 \n", + "7 42 277 282 92 \n", + "8 42 277 282 92 \n", + "9 42 277 282 92 \n", + "10 42 277 282 92 \n", + "11 42 277 282 92 \n", + "12 42 277 282 92 \n", + "13 42 277 282 92 \n", "\n", - " sample_train_10 sample_train_11 sample_train_12 sample_train_13 \\\n", - "0 148 211 60 218 \n", - "1 148 211 60 218 \n", - "2 148 211 60 218 \n", - "3 148 211 60 218 \n", - "4 148 211 60 218 \n", - ".. ... ... ... ... \n", - "275 148 211 60 218 \n", - "276 148 211 60 218 \n", - "277 148 211 60 218 \n", - "278 148 211 60 218 \n", - "279 148 211 60 218 \n", + " sample_train_10 sample_train_11 sample_train_12 sample_train_13 \\\n", + "0 148 211 60 218 \n", + "1 148 211 60 218 \n", + "2 148 211 60 218 \n", + "3 148 211 60 218 \n", + "4 148 211 60 218 \n", + "5 148 211 60 218 \n", + "6 148 211 60 218 \n", + "7 148 211 60 218 \n", + "8 148 211 60 218 \n", + "9 148 211 60 218 \n", + "10 148 211 60 218 \n", + "11 148 211 60 218 \n", + "12 148 211 60 218 \n", + "13 148 211 60 218 \n", "\n", - " sample_train_14 sample_train_15 sample_train_16 sample_train_17 \\\n", - "0 262 46 45 236 \n", - "1 262 46 45 236 \n", - "2 262 46 45 236 \n", - "3 262 46 45 236 \n", - "4 262 46 45 236 \n", - ".. ... ... ... ... \n", - "275 262 46 45 236 \n", - "276 262 46 45 236 \n", - "277 262 46 45 236 \n", - "278 262 46 45 236 \n", - "279 262 46 45 236 \n", + " sample_train_14 sample_train_15 sample_train_16 sample_train_17 \\\n", + "0 262 46 45 236 \n", + "1 262 46 45 236 \n", + "2 262 46 45 236 \n", + "3 262 46 45 236 \n", + "4 262 46 45 236 \n", + "5 262 46 45 236 \n", + "6 262 46 45 236 \n", + "7 262 46 45 236 \n", + "8 262 46 45 236 \n", + "9 262 46 45 236 \n", + "10 262 46 45 236 \n", + "11 262 46 45 236 \n", + "12 262 46 45 236 \n", + "13 262 46 45 236 \n", "\n", - " sample_train_18 sample_train_19 sample_train_20 sample_train_21 \\\n", - "0 228 132 143 167 \n", - "1 228 132 143 167 \n", - "2 228 132 143 167 \n", - "3 228 132 143 167 \n", - "4 228 132 143 167 \n", - ".. ... ... ... ... \n", - "275 228 132 143 167 \n", - "276 228 132 143 167 \n", - "277 228 132 143 167 \n", - "278 228 132 143 167 \n", - "279 228 132 143 167 \n", + " sample_train_18 sample_train_19 sample_train_20 sample_train_21 \\\n", + "0 228 132 143 167 \n", + "1 228 132 143 167 \n", + "2 228 132 143 167 \n", + "3 228 132 143 167 \n", + "4 228 132 143 167 \n", + "5 228 132 143 167 \n", + "6 228 132 143 167 \n", + "7 228 132 143 167 \n", + "8 228 132 143 167 \n", + "9 228 132 143 167 \n", + "10 228 132 143 167 \n", + "11 228 132 143 167 \n", + "12 228 132 143 167 \n", + "13 228 132 143 167 \n", "\n", - " sample_train_22 sample_train_23 sample_train_24 sample_train_25 \\\n", - "0 152 93 113 5 \n", - "1 152 93 113 5 \n", - "2 152 93 113 5 \n", - "3 152 93 113 5 \n", - "4 152 93 113 5 \n", - ".. ... ... ... ... \n", - "275 152 93 113 5 \n", - "276 152 93 113 5 \n", - "277 152 93 113 5 \n", - "278 152 93 113 5 \n", - "279 152 93 113 5 \n", + " sample_train_22 sample_train_23 sample_train_24 sample_train_25 \\\n", + "0 152 93 113 5 \n", + "1 152 93 113 5 \n", + "2 152 93 113 5 \n", + "3 152 93 113 5 \n", + "4 152 93 113 5 \n", + "5 152 93 113 5 \n", + "6 152 93 113 5 \n", + "7 152 93 113 5 \n", + "8 152 93 113 5 \n", + "9 152 93 113 5 \n", + "10 152 93 113 5 \n", + "11 152 93 113 5 \n", + "12 152 93 113 5 \n", + "13 152 93 113 5 \n", "\n", - " sample_train_26 sample_train_27 sample_train_28 sample_train_29 \\\n", - "0 238 251 170 186 \n", - "1 238 251 170 186 \n", - "2 238 251 170 186 \n", - "3 238 251 170 186 \n", - "4 238 251 170 186 \n", - ".. ... ... ... ... \n", - "275 238 251 170 186 \n", - "276 238 251 170 186 \n", - "277 238 251 170 186 \n", - "278 238 251 170 186 \n", - "279 238 251 170 186 \n", + " sample_train_26 sample_train_27 sample_train_28 sample_train_29 \\\n", + "0 238 251 170 186 \n", + "1 238 251 170 186 \n", + "2 238 251 170 186 \n", + "3 238 251 170 186 \n", + "4 238 251 170 186 \n", + "5 238 251 170 186 \n", + "6 238 251 170 186 \n", + "7 238 251 170 186 \n", + "8 238 251 170 186 \n", + "9 238 251 170 186 \n", + "10 238 251 170 186 \n", + "11 238 251 170 186 \n", + "12 238 251 170 186 \n", + "13 238 251 170 186 \n", "\n", - " sample_train_30 sample_train_31 sample_train_32 sample_train_33 \\\n", - "0 193 33 222 216 \n", - "1 193 33 222 216 \n", - "2 193 33 222 216 \n", - "3 193 33 222 216 \n", - "4 193 33 222 216 \n", - ".. ... ... ... ... \n", - "275 193 33 222 216 \n", - "276 193 33 222 216 \n", - "277 193 33 222 216 \n", - "278 193 33 222 216 \n", - "279 193 33 222 216 \n", + " sample_train_30 sample_train_31 sample_train_32 sample_train_33 \\\n", + "0 193 33 222 216 \n", + "1 193 33 222 216 \n", + "2 193 33 222 216 \n", + "3 193 33 222 216 \n", + "4 193 33 222 216 \n", + "5 193 33 222 216 \n", + "6 193 33 222 216 \n", + "7 193 33 222 216 \n", + "8 193 33 222 216 \n", + "9 193 33 222 216 \n", + "10 193 33 222 216 \n", + "11 193 33 222 216 \n", + "12 193 33 222 216 \n", + "13 193 33 222 216 \n", "\n", - " sample_train_34 sample_train_35 sample_train_36 sample_train_37 \\\n", - "0 197 73 182 119 \n", - "1 197 73 182 119 \n", - "2 197 73 182 119 \n", - "3 197 73 182 119 \n", - "4 197 73 182 119 \n", - ".. ... ... ... ... \n", - "275 197 73 182 119 \n", - "276 197 73 182 119 \n", - "277 197 73 182 119 \n", - "278 197 73 182 119 \n", - "279 197 73 182 119 \n", + " sample_train_34 sample_train_35 sample_train_36 sample_train_37 \\\n", + "0 197 73 182 119 \n", + "1 197 73 182 119 \n", + "2 197 73 182 119 \n", + "3 197 73 182 119 \n", + "4 197 73 182 119 \n", + "5 197 73 182 119 \n", + "6 197 73 182 119 \n", + "7 197 73 182 119 \n", + "8 197 73 182 119 \n", + "9 197 73 182 119 \n", + "10 197 73 182 119 \n", + "11 197 73 182 119 \n", + "12 197 73 182 119 \n", + "13 197 73 182 119 \n", "\n", - " sample_train_38 sample_train_39 sample_train_40 sample_train_41 \\\n", - "0 285 202 204 179 \n", - "1 285 202 204 179 \n", - "2 285 202 204 179 \n", - "3 285 202 204 179 \n", - "4 285 202 204 179 \n", - ".. ... ... ... ... \n", - "275 285 202 204 179 \n", - "276 285 202 204 179 \n", - "277 285 202 204 179 \n", - "278 285 202 204 179 \n", - "279 285 202 204 179 \n", + " sample_train_38 sample_train_39 sample_train_40 sample_train_41 \\\n", + "0 285 202 204 179 \n", + "1 285 202 204 179 \n", + "2 285 202 204 179 \n", + "3 285 202 204 179 \n", + "4 285 202 204 179 \n", + "5 285 202 204 179 \n", + "6 285 202 204 179 \n", + "7 285 202 204 179 \n", + "8 285 202 204 179 \n", + "9 285 202 204 179 \n", + "10 285 202 204 179 \n", + "11 285 202 204 179 \n", + "12 285 202 204 179 \n", + "13 285 202 204 179 \n", "\n", - " sample_train_42 sample_train_43 sample_train_44 sample_train_45 \\\n", - "0 177 111 59 226 \n", - "1 177 111 59 226 \n", - "2 177 111 59 226 \n", - "3 177 111 59 226 \n", - "4 177 111 59 226 \n", - ".. ... ... ... ... \n", - "275 177 111 59 226 \n", - "276 177 111 59 226 \n", - "277 177 111 59 226 \n", - "278 177 111 59 226 \n", - "279 177 111 59 226 \n", + " sample_train_42 sample_train_43 sample_train_44 sample_train_45 \\\n", + "0 177 111 59 226 \n", + "1 177 111 59 226 \n", + "2 177 111 59 226 \n", + "3 177 111 59 226 \n", + "4 177 111 59 226 \n", + "5 177 111 59 226 \n", + "6 177 111 59 226 \n", + "7 177 111 59 226 \n", + "8 177 111 59 226 \n", + "9 177 111 59 226 \n", + "10 177 111 59 226 \n", + "11 177 111 59 226 \n", + "12 177 111 59 226 \n", + "13 177 111 59 226 \n", "\n", - " sample_train_46 sample_train_47 sample_train_48 sample_train_49 \\\n", - "0 25 77 6 175 \n", - "1 25 77 6 175 \n", - "2 25 77 6 175 \n", - "3 25 77 6 175 \n", - "4 25 77 6 175 \n", - ".. ... ... ... ... \n", - "275 25 77 6 175 \n", - "276 25 77 6 175 \n", - "277 25 77 6 175 \n", - "278 25 77 6 175 \n", - "279 25 77 6 175 \n", + " sample_train_46 sample_train_47 sample_train_48 sample_train_49 \\\n", + "0 25 77 6 175 \n", + "1 25 77 6 175 \n", + "2 25 77 6 175 \n", + "3 25 77 6 175 \n", + "4 25 77 6 175 \n", + "5 25 77 6 175 \n", + "6 25 77 6 175 \n", + "7 25 77 6 175 \n", + "8 25 77 6 175 \n", + "9 25 77 6 175 \n", + "10 25 77 6 175 \n", + "11 25 77 6 175 \n", + "12 25 77 6 175 \n", + "13 25 77 6 175 \n", "\n", - " sample_train_50 sample_train_51 sample_train_52 sample_train_53 \\\n", - "0 164 140 30 22 \n", - "1 164 140 30 22 \n", - "2 164 140 30 22 \n", - "3 164 140 30 22 \n", - "4 164 140 30 22 \n", - ".. ... ... ... ... \n", - "275 164 140 30 22 \n", - "276 164 140 30 22 \n", - "277 164 140 30 22 \n", - "278 164 140 30 22 \n", - "279 164 140 30 22 \n", + " sample_train_50 sample_train_51 sample_train_52 sample_train_53 \\\n", + "0 164 140 30 22 \n", + "1 164 140 30 22 \n", + "2 164 140 30 22 \n", + "3 164 140 30 22 \n", + "4 164 140 30 22 \n", + "5 164 140 30 22 \n", + "6 164 140 30 22 \n", + "7 164 140 30 22 \n", + "8 164 140 30 22 \n", + "9 164 140 30 22 \n", + "10 164 140 30 22 \n", + "11 164 140 30 22 \n", + "12 164 140 30 22 \n", + "13 164 140 30 22 \n", "\n", - " sample_train_54 sample_train_55 sample_train_56 sample_train_57 \\\n", - "0 245 24 56 144 \n", - "1 245 24 56 144 \n", - "2 245 24 56 144 \n", - "3 245 24 56 144 \n", - "4 245 24 56 144 \n", - ".. ... ... ... ... \n", - "275 245 24 56 144 \n", - "276 245 24 56 144 \n", - "277 245 24 56 144 \n", - "278 245 24 56 144 \n", - "279 245 24 56 144 \n", + " sample_train_54 sample_train_55 sample_train_56 sample_train_57 \\\n", + "0 245 24 56 144 \n", + "1 245 24 56 144 \n", + "2 245 24 56 144 \n", + "3 245 24 56 144 \n", + "4 245 24 56 144 \n", + "5 245 24 56 144 \n", + "6 245 24 56 144 \n", + "7 245 24 56 144 \n", + "8 245 24 56 144 \n", + "9 245 24 56 144 \n", + "10 245 24 56 144 \n", + "11 245 24 56 144 \n", + "12 245 24 56 144 \n", + "13 245 24 56 144 \n", "\n", - " sample_train_58 sample_train_59 sample_train_60 sample_train_61 \\\n", - "0 124 97 63 17 \n", - "1 124 97 63 17 \n", - "2 124 97 63 17 \n", - "3 124 97 63 17 \n", - "4 124 97 63 17 \n", - ".. ... ... ... ... \n", - "275 124 97 63 17 \n", - "276 124 97 63 17 \n", - "277 124 97 63 17 \n", - "278 124 97 63 17 \n", - "279 124 97 63 17 \n", + " sample_train_58 sample_train_59 sample_train_60 sample_train_61 \\\n", + "0 124 97 63 17 \n", + "1 124 97 63 17 \n", + "2 124 97 63 17 \n", + "3 124 97 63 17 \n", + "4 124 97 63 17 \n", + "5 124 97 63 17 \n", + "6 124 97 63 17 \n", + "7 124 97 63 17 \n", + "8 124 97 63 17 \n", + "9 124 97 63 17 \n", + "10 124 97 63 17 \n", + "11 124 97 63 17 \n", + "12 124 97 63 17 \n", + "13 124 97 63 17 \n", "\n", - " sample_train_62 sample_train_63 sample_train_64 sample_train_65 \\\n", - "0 215 219 183 114 \n", - "1 215 219 183 114 \n", - "2 215 219 183 114 \n", - "3 215 219 183 114 \n", - "4 215 219 183 114 \n", - ".. ... ... ... ... \n", - "275 215 219 183 114 \n", - "276 215 219 183 114 \n", - "277 215 219 183 114 \n", - "278 215 219 183 114 \n", - "279 215 219 183 114 \n", + " sample_train_62 sample_train_63 sample_train_64 sample_train_65 \\\n", + "0 215 219 183 114 \n", + "1 215 219 183 114 \n", + "2 215 219 183 114 \n", + "3 215 219 183 114 \n", + "4 215 219 183 114 \n", + "5 215 219 183 114 \n", + "6 215 219 183 114 \n", + "7 215 219 183 114 \n", + "8 215 219 183 114 \n", + "9 215 219 183 114 \n", + "10 215 219 183 114 \n", + "11 215 219 183 114 \n", + "12 215 219 183 114 \n", + "13 215 219 183 114 \n", "\n", - " sample_train_66 sample_train_67 sample_train_68 sample_train_69 \\\n", - "0 76 284 66 178 \n", - "1 76 284 66 178 \n", - "2 76 284 66 178 \n", - "3 76 284 66 178 \n", - "4 76 284 66 178 \n", - ".. ... ... ... ... \n", - "275 76 284 66 178 \n", - "276 76 284 66 178 \n", - "277 76 284 66 178 \n", - "278 76 284 66 178 \n", - "279 76 284 66 178 \n", + " sample_train_66 sample_train_67 sample_train_68 sample_train_69 \\\n", + "0 76 284 66 178 \n", + "1 76 284 66 178 \n", + "2 76 284 66 178 \n", + "3 76 284 66 178 \n", + "4 76 284 66 178 \n", + "5 76 284 66 178 \n", + "6 76 284 66 178 \n", + "7 76 284 66 178 \n", + "8 76 284 66 178 \n", + "9 76 284 66 178 \n", + "10 76 284 66 178 \n", + "11 76 284 66 178 \n", + "12 76 284 66 178 \n", + "13 76 284 66 178 \n", "\n", - " sample_train_70 sample_train_71 sample_train_72 sample_train_73 \\\n", - "0 154 75 19 108 \n", - "1 154 75 19 108 \n", - "2 154 75 19 108 \n", - "3 154 75 19 108 \n", - "4 154 75 19 108 \n", - ".. ... ... ... ... \n", - "275 154 75 19 108 \n", - "276 154 75 19 108 \n", - "277 154 75 19 108 \n", - "278 154 75 19 108 \n", - "279 154 75 19 108 \n", + " sample_train_70 sample_train_71 sample_train_72 sample_train_73 \\\n", + "0 154 75 19 108 \n", + "1 154 75 19 108 \n", + "2 154 75 19 108 \n", + "3 154 75 19 108 \n", + "4 154 75 19 108 \n", + "5 154 75 19 108 \n", + "6 154 75 19 108 \n", + "7 154 75 19 108 \n", + "8 154 75 19 108 \n", + "9 154 75 19 108 \n", + "10 154 75 19 108 \n", + "11 154 75 19 108 \n", + "12 154 75 19 108 \n", + "13 154 75 19 108 \n", "\n", - " sample_test_0 sample_test_1 sample_test_2 sample_test_3 \\\n", - "0 69 30 39 2 \n", - "1 69 30 39 2 \n", - "2 69 30 39 2 \n", - "3 69 30 39 2 \n", - "4 69 30 39 2 \n", - ".. ... ... ... ... \n", - "275 69 30 39 2 \n", - "276 69 30 39 2 \n", - "277 69 30 39 2 \n", - "278 69 30 39 2 \n", - "279 69 30 39 2 \n", + " sample_test_0 sample_test_1 sample_test_2 sample_test_3 sample_test_4 \\\n", + "0 69 30 39 2 124 \n", + "1 69 30 39 2 124 \n", + "2 69 30 39 2 124 \n", + "3 69 30 39 2 124 \n", + "4 69 30 39 2 124 \n", + "5 69 30 39 2 124 \n", + "6 69 30 39 2 124 \n", + "7 69 30 39 2 124 \n", + "8 69 30 39 2 124 \n", + "9 69 30 39 2 124 \n", + "10 69 30 39 2 124 \n", + "11 69 30 39 2 124 \n", + "12 69 30 39 2 124 \n", + "13 69 30 39 2 124 \n", "\n", - " sample_test_4 sample_test_5 sample_test_6 sample_test_7 \\\n", - "0 124 10 68 51 \n", - "1 124 10 68 51 \n", - "2 124 10 68 51 \n", - "3 124 10 68 51 \n", - "4 124 10 68 51 \n", - ".. ... ... ... ... \n", - "275 124 10 68 51 \n", - "276 124 10 68 51 \n", - "277 124 10 68 51 \n", - "278 124 10 68 51 \n", - "279 124 10 68 51 \n", + " sample_test_5 sample_test_6 sample_test_7 sample_test_8 sample_test_9 \\\n", + "0 10 68 51 71 77 \n", + "1 10 68 51 71 77 \n", + "2 10 68 51 71 77 \n", + "3 10 68 51 71 77 \n", + "4 10 68 51 71 77 \n", + "5 10 68 51 71 77 \n", + "6 10 68 51 71 77 \n", + "7 10 68 51 71 77 \n", + "8 10 68 51 71 77 \n", + "9 10 68 51 71 77 \n", + "10 10 68 51 71 77 \n", + "11 10 68 51 71 77 \n", + "12 10 68 51 71 77 \n", + "13 10 68 51 71 77 \n", "\n", - " sample_test_8 sample_test_9 sample_test_10 sample_test_11 \\\n", - "0 71 77 102 80 \n", - "1 71 77 102 80 \n", - "2 71 77 102 80 \n", - "3 71 77 102 80 \n", - "4 71 77 102 80 \n", - ".. ... ... ... ... \n", - "275 71 77 102 80 \n", - "276 71 77 102 80 \n", - "277 71 77 102 80 \n", - "278 71 77 102 80 \n", - "279 71 77 102 80 \n", + " sample_test_10 sample_test_11 sample_test_12 sample_test_13 \\\n", + "0 102 80 76 142 \n", + "1 102 80 76 142 \n", + "2 102 80 76 142 \n", + "3 102 80 76 142 \n", + "4 102 80 76 142 \n", + "5 102 80 76 142 \n", + "6 102 80 76 142 \n", + "7 102 80 76 142 \n", + "8 102 80 76 142 \n", + "9 102 80 76 142 \n", + "10 102 80 76 142 \n", + "11 102 80 76 142 \n", + "12 102 80 76 142 \n", + "13 102 80 76 142 \n", "\n", - " sample_test_12 sample_test_13 sample_test_14 sample_test_15 \\\n", - "0 76 142 127 95 \n", - "1 76 142 127 95 \n", - "2 76 142 127 95 \n", - "3 76 142 127 95 \n", - "4 76 142 127 95 \n", - ".. ... ... ... ... \n", - "275 76 142 127 95 \n", - "276 76 142 127 95 \n", - "277 76 142 127 95 \n", - "278 76 142 127 95 \n", - "279 76 142 127 95 \n", + " sample_test_14 sample_test_15 sample_test_16 sample_test_17 \\\n", + "0 127 95 70 93 \n", + "1 127 95 70 93 \n", + "2 127 95 70 93 \n", + "3 127 95 70 93 \n", + "4 127 95 70 93 \n", + "5 127 95 70 93 \n", + "6 127 95 70 93 \n", + "7 127 95 70 93 \n", + "8 127 95 70 93 \n", + "9 127 95 70 93 \n", + "10 127 95 70 93 \n", + "11 127 95 70 93 \n", + "12 127 95 70 93 \n", + "13 127 95 70 93 \n", "\n", - " sample_test_16 sample_test_17 sample_test_18 sample_test_19 \\\n", - "0 70 93 67 0 \n", - "1 70 93 67 0 \n", - "2 70 93 67 0 \n", - "3 70 93 67 0 \n", - "4 70 93 67 0 \n", - ".. ... ... ... ... \n", - "275 70 93 67 0 \n", - "276 70 93 67 0 \n", - "277 70 93 67 0 \n", - "278 70 93 67 0 \n", - "279 70 93 67 0 \n", + " sample_test_18 sample_test_19 sample_test_20 sample_test_21 \\\n", + "0 67 0 105 82 \n", + "1 67 0 105 82 \n", + "2 67 0 105 82 \n", + "3 67 0 105 82 \n", + "4 67 0 105 82 \n", + "5 67 0 105 82 \n", + "6 67 0 105 82 \n", + "7 67 0 105 82 \n", + "8 67 0 105 82 \n", + "9 67 0 105 82 \n", + "10 67 0 105 82 \n", + "11 67 0 105 82 \n", + "12 67 0 105 82 \n", + "13 67 0 105 82 \n", "\n", - " sample_test_20 sample_test_21 sample_test_22 sample_test_23 \\\n", - "0 105 82 136 40 \n", - "1 105 82 136 40 \n", - "2 105 82 136 40 \n", - "3 105 82 136 40 \n", - "4 105 82 136 40 \n", - ".. ... ... ... ... \n", - "275 105 82 136 40 \n", - "276 105 82 136 40 \n", - "277 105 82 136 40 \n", - "278 105 82 136 40 \n", - "279 105 82 136 40 \n", + " sample_test_22 sample_test_23 sample_test_24 sample_test_25 \\\n", + "0 136 40 54 28 \n", + "1 136 40 54 28 \n", + "2 136 40 54 28 \n", + "3 136 40 54 28 \n", + "4 136 40 54 28 \n", + "5 136 40 54 28 \n", + "6 136 40 54 28 \n", + "7 136 40 54 28 \n", + "8 136 40 54 28 \n", + "9 136 40 54 28 \n", + "10 136 40 54 28 \n", + "11 136 40 54 28 \n", + "12 136 40 54 28 \n", + "13 136 40 54 28 \n", "\n", - " sample_test_24 sample_test_25 sample_test_26 sample_test_27 \\\n", - "0 54 28 74 119 \n", - "1 54 28 74 119 \n", - "2 54 28 74 119 \n", - "3 54 28 74 119 \n", - "4 54 28 74 119 \n", - ".. ... ... ... ... \n", - "275 54 28 74 119 \n", - "276 54 28 74 119 \n", - "277 54 28 74 119 \n", - "278 54 28 74 119 \n", - "279 54 28 74 119 \n", + " sample_test_26 sample_test_27 sample_test_28 sample_test_29 \\\n", + "0 74 119 18 9 \n", + "1 74 119 18 9 \n", + "2 74 119 18 9 \n", + "3 74 119 18 9 \n", + "4 74 119 18 9 \n", + "5 74 119 18 9 \n", + "6 74 119 18 9 \n", + "7 74 119 18 9 \n", + "8 74 119 18 9 \n", + "9 74 119 18 9 \n", + "10 74 119 18 9 \n", + "11 74 119 18 9 \n", + "12 74 119 18 9 \n", + "13 74 119 18 9 \n", "\n", - " sample_test_28 sample_test_29 sample_test_30 sample_test_31 \\\n", - "0 18 9 58 99 \n", - "1 18 9 58 99 \n", - "2 18 9 58 99 \n", - "3 18 9 58 99 \n", - "4 18 9 58 99 \n", - ".. ... ... ... ... \n", - "275 18 9 58 99 \n", - "276 18 9 58 99 \n", - "277 18 9 58 99 \n", - "278 18 9 58 99 \n", - "279 18 9 58 99 \n", + " sample_test_30 sample_test_31 sample_test_32 sample_test_33 \\\n", + "0 58 99 73 97 \n", + "1 58 99 73 97 \n", + "2 58 99 73 97 \n", + "3 58 99 73 97 \n", + "4 58 99 73 97 \n", + "5 58 99 73 97 \n", + "6 58 99 73 97 \n", + "7 58 99 73 97 \n", + "8 58 99 73 97 \n", + "9 58 99 73 97 \n", + "10 58 99 73 97 \n", + "11 58 99 73 97 \n", + "12 58 99 73 97 \n", + "13 58 99 73 97 \n", "\n", - " sample_test_32 sample_test_33 sample_test_34 sample_test_35 \\\n", - "0 73 97 128 122 \n", - "1 73 97 128 122 \n", - "2 73 97 128 122 \n", - "3 73 97 128 122 \n", - "4 73 97 128 122 \n", - ".. ... ... ... ... \n", - "275 73 97 128 122 \n", - "276 73 97 128 122 \n", - "277 73 97 128 122 \n", - "278 73 97 128 122 \n", - "279 73 97 128 122 \n", + " sample_test_34 sample_test_35 fi_time train_AUROC train_AUPRC \\\n", + "0 128 122 157.653250 0.884865 0.907278 \n", + "1 128 122 2.366904 0.590270 0.731816 \n", + "2 128 122 2.749968 0.574054 0.722772 \n", + "3 128 122 66.605561 0.581081 0.729977 \n", + "4 128 122 4.271746 0.537297 0.667003 \n", + "5 128 122 170.402314 0.876216 0.905993 \n", + "6 128 122 0.414457 0.805946 0.821571 \n", + "7 128 122 158.937756 0.954595 0.967038 \n", + "8 128 122 2.351976 0.643243 0.788523 \n", + "9 128 122 2.728971 0.637838 0.785668 \n", + "10 128 122 63.704902 0.650811 0.798041 \n", + "11 128 122 4.068172 0.550811 0.684134 \n", + "12 128 122 172.658590 0.984865 0.986969 \n", + "13 128 122 0.400583 0.793514 0.831404 \n", "\n", - " fi_time train_AUROC train_AUPRC train_F1 test_AUROC test_AUPRC \\\n", - "0 1622.727949 0.577838 0.635698 0.000000 0.646667 0.677123 \n", - "1 11.966596 0.495676 0.612882 0.130066 0.451111 0.581673 \n", - "2 12.777327 0.474595 0.598768 0.149544 0.451111 0.581673 \n", - "3 0.000003 0.521622 0.638607 0.101366 0.476667 0.588973 \n", - "4 23.028107 0.521622 0.638607 0.101366 0.476667 0.588973 \n", - ".. ... ... ... ... ... ... \n", - "275 10.731668 0.327568 0.573769 0.000000 0.272222 0.543042 \n", - "276 0.000002 0.433514 0.614721 0.000000 0.410000 0.596803 \n", - "277 19.305907 0.433514 0.614721 0.000000 0.410000 0.596803 \n", - "278 1775.899545 0.946486 0.957699 0.000000 0.952222 0.962438 \n", - "279 4.426196 0.848108 0.888083 0.000000 0.828889 0.872943 \n", - "\n", - " test_F1 split_seed \n", - "0 0.000000 1 \n", - "1 0.013889 1 \n", - "2 0.013889 1 \n", - "3 0.029101 1 \n", - "4 0.029101 1 \n", - ".. ... ... \n", - "275 0.000000 3 \n", - "276 0.000000 3 \n", - "277 0.000000 3 \n", - "278 0.000000 3 \n", - "279 0.000000 3 \n", - "\n", - "[280 rows x 134 columns]" + " train_F1 test_AUROC test_AUPRC test_F1 split_seed \n", + "0 0.0 0.930000 0.941834 0.0 1 \n", + "1 0.0 0.574444 0.724134 0.0 1 \n", + "2 0.0 0.574444 0.724134 0.0 1 \n", + "3 0.0 0.597778 0.743177 0.0 1 \n", + "4 0.0 0.524444 0.662643 0.0 1 \n", + "5 0.0 0.877778 0.909645 0.0 1 \n", + "6 0.0 0.846667 0.862467 0.0 1 \n", + "7 0.0 0.983333 0.988194 0.0 1 \n", + "8 0.0 0.594444 0.759802 0.0 1 \n", + "9 0.0 0.594444 0.759802 0.0 1 \n", + "10 0.0 0.598889 0.764195 0.0 1 \n", + "11 0.0 0.521111 0.667200 0.0 1 \n", + "12 0.0 0.983333 0.986455 0.0 1 \n", + "13 0.0 0.831111 0.862800 0.0 1 " ] }, - "execution_count": 5, + "execution_count": 3, "metadata": {}, "output_type": "execute_result" } @@ -2151,7 +2642,7 @@ }, { "cell_type": "code", - "execution_count": 6, + "execution_count": 4, "metadata": {}, "outputs": [ { @@ -2169,7 +2660,7 @@ }, { "cell_type": "code", - "execution_count": 8, + "execution_count": 5, "metadata": {}, "outputs": [], "source": [ @@ -2197,7 +2688,7 @@ }, { "cell_type": "code", - "execution_count": 9, + "execution_count": 6, "metadata": {}, "outputs": [], "source": [ @@ -2223,7 +2714,7 @@ }, { "cell_type": "code", - "execution_count": 10, + "execution_count": 7, "metadata": {}, "outputs": [], "source": [ @@ -2242,7 +2733,7 @@ }, { "cell_type": "code", - "execution_count": 11, + "execution_count": 8, "metadata": {}, "outputs": [], "source": [ @@ -2288,183 +2779,99 @@ }, { "cell_type": "code", - "execution_count": 16, + "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/plain": [ - "{0.1: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.5492222222222223],\n", - " 'train_auroc': [0.5481621621621622],\n", - " 'test_auprc': [0.6498311287477954],\n", - " 'train_auprc': [0.6492926855426855],\n", - " 'test_f1': [0.00277777777777777],\n", - " 'train_f1': [0.00315315315315315]},\n", - " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.3757777777777777],\n", - " 'train_auroc': [0.43264864864864866],\n", - " 'test_auprc': [0.5232012786596119],\n", - " 'train_auprc': [0.5762566495066495],\n", - " 'test_f1': [0.01772425105758434],\n", - " 'train_f1': [0.16920817920817915]},\n", - " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.3757777777777777],\n", - " 'train_auroc': [0.438054054054054],\n", - " 'test_auprc': [0.5232012786596119],\n", - " 'train_auprc': [0.5781234448734448],\n", - " 'test_f1': [0.01772425105758434],\n", - " 'train_f1': [0.18028579528579525]},\n", - " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.4962222222222222],\n", - " 'train_auroc': [0.5052972972972973],\n", - " 'test_auprc': [0.6122768959435626],\n", - " 'train_auprc': [0.6334653582153582],\n", - " 'test_f1': [0.02094029427362757],\n", - " 'train_f1': [0.14491034491034482]},\n", - " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.4962222222222222],\n", - " 'train_auroc': [0.5052972972972973],\n", - " 'test_auprc': [0.6122768959435626],\n", - " 'train_auprc': [0.6334653582153582],\n", - " 'test_f1': [0.02094029427362757],\n", - " 'train_f1': [0.14491034491034482]},\n", - " 'LIME_RF_plus': {'test_auroc': [0.5762222222222223],\n", - " 'train_auroc': [0.5754594594594595],\n", - " 'test_auprc': [0.6635224867724867],\n", - " 'train_auprc': [0.667269734019734],\n", + "{0.4: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.09300000000000001],\n", + " 'train_auroc': [0.0884864864864865],\n", + " 'test_auprc': [0.09418342151675484],\n", + " 'train_auprc': [0.09072779922779925],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'TreeSHAP_RF': {'test_auroc': [0.6147777777777778],\n", - " 'train_auroc': [0.5977297297297298],\n", - " 'test_auprc': [0.680263888888889],\n", - " 'train_auprc': [0.6812614757614757],\n", - " 'test_f1': [0.00092592592592592],\n", - " 'train_f1': [0.0]}},\n", - " 0.2: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.6562222222222223],\n", - " 'train_auroc': [0.6285945945945945],\n", - " 'test_auprc': [0.7265101410934743],\n", - " 'train_auprc': [0.709574217074217],\n", - " 'test_f1': [0.00092592592592592],\n", - " 'train_f1': [0.0]},\n", - " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.35044444444444445],\n", - " 'train_auroc': [0.41248648648648645],\n", - " 'test_auprc': [0.5209351851851852],\n", - " 'train_auprc': [0.5772712355212355],\n", - " 'test_f1': [0.00131172839506172],\n", - " 'train_f1': [0.07541244041244036]},\n", - " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.35044444444444445],\n", - " 'train_auroc': [0.41291891891891896],\n", - " 'test_auprc': [0.5209351851851852],\n", - " 'train_auprc': [0.5778686186186186],\n", - " 'test_f1': [0.00131172839506172],\n", - " 'train_f1': [0.08324372324372317]},\n", - " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.49822222222222223],\n", - " 'train_auroc': [0.4912432432432432],\n", - " 'test_auprc': [0.6218267195767195],\n", - " 'train_auprc': [0.6302449592449592],\n", - " 'test_f1': [0.0],\n", - " 'train_f1': [0.06847857097857092]},\n", - " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.49822222222222223],\n", - " 'train_auroc': [0.4912432432432432],\n", - " 'test_auprc': [0.6218267195767195],\n", - " 'train_auprc': [0.6302449592449592],\n", - " 'test_f1': [0.0],\n", - " 'train_f1': [0.06847857097857092]},\n", - " 'LIME_RF_plus': {'test_auroc': [0.6900000000000002],\n", - " 'train_auroc': [0.6841621621621622],\n", - " 'test_auprc': [0.742490520282187],\n", - " 'train_auprc': [0.7372000214500214],\n", + " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.057444444444444444],\n", + " 'train_auroc': [0.059027027027027036],\n", + " 'test_auprc': [0.07241335978835978],\n", + " 'train_auprc': [0.07318157443157443],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'TreeSHAP_RF': {'test_auroc': [0.709],\n", - " 'train_auroc': [0.6683783783783784],\n", - " 'test_auprc': [0.7508511904761905],\n", - " 'train_auprc': [0.7326668811668811],\n", - " 'test_f1': [0.0],\n", - " 'train_f1': [0.0]}},\n", - " 0.4: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.8031111111111112],\n", - " 'train_auroc': [0.7802702702702702],\n", - " 'test_auprc': [0.8474830246913581],\n", - " 'train_auprc': [0.8309603174603174],\n", + " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.057444444444444444],\n", + " 'train_auroc': [0.05740540540540541],\n", + " 'test_auprc': [0.07241335978835978],\n", + " 'train_auprc': [0.07227724152724153],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.33299999999999996],\n", - " 'train_auroc': [0.3830810810810811],\n", - " 'test_auprc': [0.5429991181657849],\n", - " 'train_auprc': [0.5844474474474474],\n", + " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.059777777777777784],\n", + " 'train_auroc': [0.058108108108108104],\n", + " 'test_auprc': [0.07431768077601411],\n", + " 'train_auprc': [0.07299774774774774],\n", " 'test_f1': [0.0],\n", - " 'train_f1': [0.013756558756558717]},\n", - " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.33299999999999996],\n", - " 'train_auroc': [0.3847027027027027],\n", - " 'test_auprc': [0.5429991181657849],\n", - " 'train_auprc': [0.5825757185757185],\n", - " 'test_f1': [0.0],\n", - " 'train_f1': [0.015869565869565817]},\n", - " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.4681111111111111],\n", - " 'train_auroc': [0.46075675675675676],\n", - " 'test_auprc': [0.6163648589065256],\n", - " 'train_auprc': [0.6209760832260832],\n", - " 'test_f1': [0.0],\n", - " 'train_f1': [0.011945329445329401]},\n", - " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.4681111111111111],\n", - " 'train_auroc': [0.46075675675675676],\n", - " 'test_auprc': [0.6163648589065256],\n", - " 'train_auprc': [0.6209760832260832],\n", + " 'train_f1': [0.0]},\n", + " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.05244444444444444],\n", + " 'train_auroc': [0.053729729729729725],\n", + " 'test_auprc': [0.06626432980599646],\n", + " 'train_auprc': [0.0667003432003432],\n", " 'test_f1': [0.0],\n", - " 'train_f1': [0.011945329445329401]},\n", - " 'LIME_RF_plus': {'test_auroc': [0.841777777777778],\n", - " 'train_auroc': [0.8435675675675676],\n", - " 'test_auprc': [0.8751393298059966],\n", - " 'train_auprc': [0.8756224796224796],\n", + " 'train_f1': [0.0]},\n", + " 'LIME_RF_plus': {'test_auroc': [0.08777777777777779],\n", + " 'train_auroc': [0.08762162162162163],\n", + " 'test_auprc': [0.0909645061728395],\n", + " 'train_auprc': [0.09059931359931359],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'TreeSHAP_RF': {'test_auroc': [0.7861111111111112],\n", - " 'train_auroc': [0.760972972972973],\n", - " 'test_auprc': [0.8175163139329804],\n", - " 'train_auprc': [0.8062584727584727],\n", + " 'TreeSHAP_RF': {'test_auroc': [0.0846666666666667],\n", + " 'train_auroc': [0.08059459459459459],\n", + " 'test_auprc': [0.08624669312169311],\n", + " 'train_auprc': [0.08215712140712142],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]}},\n", - " 0.8: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.9321111111111111],\n", - " 'train_auroc': [0.9180270270270272],\n", - " 'test_auprc': [0.9542843915343916],\n", - " 'train_auprc': [0.9443765551265552],\n", + " 0.8: {'Kernel_SHAP_RF_plus': {'test_auroc': [0.09833333333333331],\n", + " 'train_auroc': [0.09545945945945944],\n", + " 'test_auprc': [0.09881944444444443],\n", + " 'train_auprc': [0.0967037752037752],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.32611111111111113],\n", - " 'train_auroc': [0.34589189189189196],\n", - " 'test_auprc': [0.5782162698412698],\n", - " 'train_auprc': [0.5901048906048906],\n", + " 'LFI_evaluate_on_all_RF_plus': {'test_auroc': [0.05944444444444445],\n", + " 'train_auroc': [0.0643243243243243],\n", + " 'test_auprc': [0.0759801587301587],\n", + " 'train_auprc': [0.07885231660231659],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.32611111111111113],\n", - " 'train_auroc': [0.34627027027027024],\n", - " 'test_auprc': [0.5782162698412698],\n", - " 'train_auprc': [0.5884340411840411],\n", + " 'LFI_evaluate_on_oob_RF_plus': {'test_auroc': [0.05944444444444445],\n", + " 'train_auroc': [0.06378378378378378],\n", + " 'test_auprc': [0.0759801587301587],\n", + " 'train_auprc': [0.07856681681681682],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.4306666666666666],\n", - " 'train_auroc': [0.4362702702702704],\n", - " 'test_auprc': [0.614150352733686],\n", - " 'train_auprc': [0.6165018232518232],\n", + " 'LFI_fit_on_OOB_RF': {'test_auroc': [0.059888888888888894],\n", + " 'train_auroc': [0.06508108108108109],\n", + " 'test_auprc': [0.07641953262786595],\n", + " 'train_auprc': [0.07980405405405405],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.4306666666666666],\n", - " 'train_auroc': [0.4362702702702704],\n", - " 'test_auprc': [0.614150352733686],\n", - " 'train_auprc': [0.6165018232518232],\n", + " 'LFI_fit_on_inbag_RF': {'test_auroc': [0.05211111111111112],\n", + " 'train_auroc': [0.055081081081081076],\n", + " 'test_auprc': [0.06672001763668431],\n", + " 'train_auprc': [0.06841344916344917],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'LIME_RF_plus': {'test_auroc': [0.9651111111111111],\n", - " 'train_auroc': [0.9619459459459458],\n", - " 'test_auprc': [0.9720612874779542],\n", - " 'train_auprc': [0.9693102745602745],\n", + " 'LIME_RF_plus': {'test_auroc': [0.09833333333333331],\n", + " 'train_auroc': [0.09848648648648647],\n", + " 'test_auprc': [0.09864550264550263],\n", + " 'train_auprc': [0.0986969111969112],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]},\n", - " 'TreeSHAP_RF': {'test_auroc': [0.8405555555555555],\n", - " 'train_auroc': [0.835135135135135],\n", - " 'test_auprc': [0.8785090388007054],\n", - " 'train_auprc': [0.8749084084084083],\n", + " 'TreeSHAP_RF': {'test_auroc': [0.08311111111111112],\n", + " 'train_auroc': [0.07935135135135134],\n", + " 'test_auprc': [0.0862799823633157],\n", + " 'train_auprc': [0.08314039039039038],\n", " 'test_f1': [0.0],\n", " 'train_f1': [0.0]}}}" ] }, - "execution_count": 16, + "execution_count": 9, "metadata": {}, "output_type": "execute_result" } @@ -2475,14 +2882,14 @@ }, { "cell_type": "code", - "execution_count": 18, + "execution_count": 19, "metadata": {}, "outputs": [ { "data": { - "image/png": "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", + "image/png": "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", "text/plain": [ - "
" + "
" ] }, "metadata": {}, @@ -2491,14 +2898,13 @@ ], "source": [ "# create barplot with four subplots, one for each heritability level, make x axis text vertical\n", - "fig, axs = plt.subplots(2, 2, figsize=(12, 12))\n", + "fig, axs = plt.subplots(1, 2, figsize=(12, 12))\n", "fig.suptitle(\"Performance of different feature importance methods on diabetes regression task with varying heritability levels\")\n", "for i, h in enumerate(heritability):\n", " for j, m in enumerate(methods):\n", - " axs[i//2, i%2].bar(m, agg_results[h][m][\"test_auroc\"][0], label=f\"{m}\", alpha=0.7)\n", - " axs[i//2, i%2].set_title(f\"Heritability = {h}\")\n", - " axs[i//2, i%2].set_ylabel(\"Test AUROC\")\n", - " axs[i//2, i%2].set_ylim(0.5, 1)\n", + " axs[i].bar(m, agg_results[h][m][\"test_auroc\"][0], label=f\"{m}\", alpha=0.7)\n", + " axs[i].set_title(f\"Heritability = {h}\")\n", + " axs[i].set_ylabel(\"Test AUROC\")\n", " plt.xticks(rotation=45)" ] }, diff --git a/feature_importance/ranking_importance_local_sims.py b/feature_importance/ranking_importance_local_sims.py index 323941f..15874d6 100644 --- a/feature_importance/ranking_importance_local_sims.py +++ b/feature_importance/ranking_importance_local_sims.py @@ -96,13 +96,6 @@ def compare_estimators(estimators: List[ModelConfig], # X = normalizer.fit_transform(X) # X_train = normalizer.transform(X_train) # X_test = normalizer.transform(X_test) - - print("Line 85") - - # fit model - est.fit(X_train, y_train) - - print("Line 90") np.random.seed(42) indices_train = np.random.choice(X_train.shape[0], int(X_train.shape[0]*.25), replace=False) @@ -139,7 +132,7 @@ def compare_estimators(estimators: List[ModelConfig], fit=rf_plus_base, **fi_est.kwargs) local_fi_score_train_subset = local_fi_score_train[indices_train] local_partial_pred_train_subset = local_parital_pred_train[indices_train] - elif fi_est.name == "LFI_fit_on_inbag_RF" or fi_est.name == "LFI_fit_on_inbag_RF": + elif fi_est.name == "LFI_fit_on_inbag_RF" or fi_est.name == "LFI_fit_on_OOB_RF": local_fi_score_train, local_parital_pred_train, local_fi_score_test, local_partial_pred_test, local_fi_score_test_subset, local_partial_pred_test_subset = fi_est.cls(X_train=X_train, y_train=y_train, X_train_subset = X_train_subset, y_train_subset=y_train_subset, X_test_subset=X_test_subset, X_test=X_test, diff --git a/feature_importance/scripts/competing_methods_local.py b/feature_importance/scripts/competing_methods_local.py index 999c823..6894e6a 100644 --- a/feature_importance/scripts/competing_methods_local.py +++ b/feature_importance/scripts/competing_methods_local.py @@ -67,10 +67,10 @@ def LFI_evaluation_RF_MDI(X_train, y_train, X_train_subset, y_train_subset, X_te rf_plus_mdi = RFPlusMDI(rf_plus_model, evaluate_on="all") num_samples, num_features = X_data.shape local_feature_importances, partial_preds = rf_plus_mdi.explain(X=X_data, y=y_data) - abs_local_feature_importances = np.abs(local_feature_importances) - abs_partial_preds = np.abs(partial_preds) - result_tables.append(abs_local_feature_importances) - result_tables.append(abs_partial_preds) + # abs_local_feature_importances = np.abs(local_feature_importances) + # abs_partial_preds = np.abs(partial_preds) + result_tables.append(local_feature_importances) # used to be abs + result_tables.append(partial_preds) # used to be abs return tuple(result_tables) def LFI_evaluation_RF_MDI_classification(X_train, y_train, X_train_subset, y_train_subset, X_test, X_test_subset, fit, **kwargs): @@ -90,10 +90,10 @@ def LFI_evaluation_RF_MDI_classification(X_train, y_train, X_train_subset, y_tra rf_plus_mdi = RFPlusMDI(rf_plus_model, evaluate_on="all") num_samples, num_features = X_data.shape local_feature_importances, partial_preds = rf_plus_mdi.explain(X=X_data, y=y_data) - abs_local_feature_importances = np.abs(local_feature_importances) - abs_partial_preds = np.abs(partial_preds) - result_tables.append(abs_local_feature_importances) - result_tables.append(abs_partial_preds) + # abs_local_feature_importances = np.abs(local_feature_importances) + # abs_partial_preds = np.abs(partial_preds) + result_tables.append(local_feature_importances) # used to be abs + result_tables.append(partial_preds) # used to be abs return tuple(result_tables) def LFI_evaluation_RF_OOB(X_train, y_train, X_train_subset, y_train_subset, X_test, X_test_subset, fit, **kwargs): @@ -117,10 +117,10 @@ def LFI_evaluation_RF_OOB(X_train, y_train, X_train_subset, y_train_subset, X_te rf_plus_mdi = AloRFPlusMDI(rf_plus_model, evaluate_on="all") num_samples, num_features = X_data.shape local_feature_importances, partial_preds = rf_plus_mdi.explain(X=X_data, y=y_data) - abs_local_feature_importances = np.abs(local_feature_importances) - abs_partial_preds = np.abs(partial_preds) - result_tables.append(abs_local_feature_importances) - result_tables.append(abs_partial_preds) + # abs_local_feature_importances = np.abs(local_feature_importances) + # abs_partial_preds = np.abs(partial_preds) + result_tables.append(local_feature_importances) # used to be abs + result_tables.append(partial_preds) # used to be abs return tuple(result_tables) @@ -134,10 +134,10 @@ def LFI_evaluation_RF_plus(X_train, y_train, X_train_subset, y_train_subset, X_t for X_data, y_data in subsets: num_samples, num_features = X_data.shape local_feature_importances, partial_preds = rf_plus_mdi.explain(X=X_data, y=y_data) - abs_local_feature_importances = np.abs(local_feature_importances) - abs_partial_preds = np.abs(partial_preds) - result_tables.append(abs_local_feature_importances) - result_tables.append(abs_partial_preds) + # abs_local_feature_importances = np.abs(local_feature_importances) + # abs_partial_preds = np.abs(partial_preds) + result_tables.append(local_feature_importances) # used to be abs + result_tables.append(partial_preds) # used to be abs return tuple(result_tables) @@ -153,10 +153,10 @@ def LFI_evaluation_RF_plus_OOB(X_train, y_train, X_train_subset, y_train_subset, else: rf_plus_mdi = AloRFPlusMDI(fit, evaluate_on="all") local_feature_importances, partial_preds = rf_plus_mdi.explain(X=X_data, y=y_data) - abs_local_feature_importances = np.abs(local_feature_importances) - abs_partial_preds = np.abs(partial_preds) - result_tables.append(abs_local_feature_importances) - result_tables.append(abs_partial_preds) + # abs_local_feature_importances = np.abs(local_feature_importances) + # abs_partial_preds = np.abs(partial_preds) + result_tables.append(local_feature_importances) # used to be abs + result_tables.append(partial_preds) # used to be abs return tuple(result_tables)