eyra · giantori · Apr 20, 2024 · Apr 20, 2024 · Apr 20, 2024 · Apr 20, 2024
diff --git a/.gitignore b/.gitignore
@@ -1,7 +1,9 @@
 .ipynb_checkpoints
+catboost_info/
+
 # macOS
 .DS_Store
 .AppleDouble
 .LSOverride
 __pycache__/
-.tool-versions
+.tool-versions
diff --git a/PreFer_codebook.csv b/PreFer_codebook.csv
diff --git a/PreFer_fake_data.csv b/PreFer_fake_data.csv
diff --git a/description.md b/description.md
diff --git a/environment.yml b/environment.yml
diff --git a/model.joblib b/model.joblib
diff --git a/model.rds b/model.rds
diff --git a/mydata.csv b/mydata.csv
diff --git a/packages.R b/packages.R
diff --git a/r.Dockerfile b/r.Dockerfile
diff --git a/run.R b/run.R
diff --git a/run.py b/run.py
@@ -22,10 +22,8 @@
 
 parser = argparse.ArgumentParser(description="Process data.")
 
-parser.add_argument("data_path", help="Path to data data CSV file.")
-parser.add_argument(
-    "background_data_path", help="Path to background data data CSV file."
-)
+parser.add_argument("data_path", help="Path to data CSV file.")
+parser.add_argument("background_data_path", help="Path to background data CSV file.")
 parser.add_argument("--output", help="Path to prediction output CSV file.")
 
 args = parser.parse_args()

diff --git a/saved/FA_loadings.png b/saved/FA_loadings.png
diff --git a/saved/FA_screeplot.png b/saved/FA_screeplot.png
diff --git a/saved/PreFer_logo.png b/saved/PreFer_logo.png
diff --git a/saved/cleaned_df.csv b/saved/cleaned_df.csv
diff --git a/saved/correlation_matrix.png b/saved/correlation_matrix.png
diff --git a/saved/eda_feature_plots.png b/saved/eda_feature_plots.png
diff --git a/saved/model_df.csv b/saved/model_df.csv
diff --git a/saved/model_df_display.csv b/saved/model_df_display.csv
diff --git a/saved/model_selection_metrics.csv b/saved/model_selection_metrics.csv
@@ -0,0 +1,16 @@
+model,accuracy,precision,recall,f1
+voting_s,0.9127946127946128,0.857655842332095,0.715625,0.7791283588871434
+voting_h,0.9148148148148149,0.882570177083562,0.6984375,0.7788772847556219
+catb,0.9141414141414141,0.876642223380349,0.7015625,0.7783646333847204
+stacking lgb-catb,0.9020202020202019,0.7850379028346589,0.753125,0.7678774865522702
+stacking et-catb-lgb,0.8993265993265993,0.7683470485290076,0.765625,0.7660368238861996
+stacking et-catb,0.8979797979797979,0.7598225487523408,0.771875,0.7650820760768711
+lgb,0.9033670033670033,0.8156425708098813,0.7140625,0.7606226540365956
+et,0.9084175084175083,0.8751459981254814,0.671875,0.7594668918530201
+xgb,0.8986531986531986,0.8134420781748741,0.6890625,0.7451552904907756
+rf,0.8905723905723905,0.8155900332360261,0.6375,0.7143824210961051
+svc,0.8764309764309764,0.8571573967159598,0.5125,0.6395062345966023
+dt,0.8400673400673402,0.6284819142456894,0.6375,0.6314023658050075
+mlp,0.8457912457912459,0.6611369798887063,0.603125,0.6284527078388321
+lr,0.7855218855218855,0.5024562468626843,0.7265625,0.5931716897153938
+knn,0.8494949494949495,0.7896225014928944,0.4171875,0.542875159699717
diff --git a/saved/scores_newchild.png b/saved/scores_newchild.png
diff --git a/saved/shap_bar.png b/saved/shap_bar.png
diff --git a/saved/shap_beeswarm.png b/saved/shap_beeswarm.png
diff --git a/saved/shap_scatter_plots.png b/saved/shap_scatter_plots.png
diff --git a/saved/shap_values_cb.joblib b/saved/shap_values_cb.joblib
diff --git a/saved/tree_selection_2020.png b/saved/tree_selection_2020.png
diff --git a/saved/tree_selection_2020_Imputed.png b/saved/tree_selection_2020_Imputed.png
diff --git a/saved/tree_selection_2020_Imputed2.png b/saved/tree_selection_2020_Imputed2.png
diff --git a/saved/tree_selection_NAimputed.png b/saved/tree_selection_NAimputed.png
diff --git a/saved/tree_selection_background.png b/saved/tree_selection_background.png
diff --git a/saved/tree_selection_big_year.png b/saved/tree_selection_big_year.png
diff --git a/saved/tree_selection_cleaned.png b/saved/tree_selection_cleaned.png