Update src/bin/binding_affinity_estimator.py

src/bin/binding_affinity_estimator.py

@@ -47,42 +47,28 @@ def calc_validation_error(X_test, y_test, model):
 def calc_metrics(X_train, y_train, X_test, y_test, model):
     '''Fits the model and returns the metrics for in-sample and out-of-sample errors.'''
     model.fit(X_train, y_train)
-    train_mse_error, train_mae_error, train_corr = calc_train_error(X_train, y_train, model)
+    #train_mse_error, train_mae_error, train_corr = calc_train_error(X_train, y_train, model)
     val_mse_error, val_mae_error, val_corr = calc_validation_error(X_test, y_test, model)
-    return train_mse_error, val_mse_error, train_mae_error, val_mae_error, train_corr, val_corr
+    return val_mse_error, val_mae_error, val_corr
 
 def report_results(
-    train_mse_error_list,
     validation_mse_error_list,
-    train_mae_error_list,
     validation_mae_error_list,
-    train_corr_list,
     validation_corr_list,
-    train_corr_pval_list,
     validation_corr_pval_list,
 ):
     result_summary = {
-        "train_mse_error": round(np.mean(train_mse_error_list) * 100, 4),
-        "train_mse_std": round(np.std(train_mse_error_list) * 100, 4),
         "val_mse_error": round(np.mean(validation_mse_error_list) * 100, 4),
         "val_mse_std": round(np.std(validation_mse_error_list) * 100, 4),
-        "train_mae_error": round(np.mean(train_mae_error_list) * 100, 4),
-        "train_mae_std": round(np.std(train_mae_error_list) * 100, 4),
         "val_mae_error": round(np.mean(validation_mae_error_list) * 100, 4),
         "val_mae_std": round(np.std(validation_mae_error_list) * 100, 4),
-        "train_corr": round(np.mean(train_corr_list), 4),
-        "train_corr_pval": round(np.mean(train_corr_pval_list), 4),
         "validation_corr": round(np.mean(validation_corr_list), 4),
         "validation_corr_pval": round(np.mean(validation_corr_pval_list), 4),
     }
 
     result_detail = {
-        "train_mse_errors": list(np.multiply(train_mse_error_list, 100)),
         "val_mse_errors": list(np.multiply(validation_mse_error_list, 100)),
-        "train_mae_errors": list(np.multiply(train_mae_error_list, 100)),
         "val_mae_errors": list(np.multiply(validation_mae_error_list, 100)),
-        "train_corrs": list(np.multiply(train_corr_list, 100)),
-        "train_corr_pvals": list(np.multiply(train_corr_pval_list, 100)),
         "validation_corrs": list(np.multiply(validation_corr_list, 100)),
         "validation_corr_pvals": list(np.multiply(validation_corr_pval_list, 100)),
     }
@@ -123,35 +109,24 @@ def predictAffinityWithModel(regressor_model, multiplied_vectors_df):
 
         # calculate errors
         (
-            train_mse_error,
             val_mse_error,
-            train_mae_error,
             val_mae_error,
-            train_corr,
             val_corr,
         ) = calc_metrics(X_train, y_train, X_val, y_val, reg)
 
         # append to appropriate lists
-        train_mse_error_list.append(train_mse_error)
         validation_mse_error_list.append(val_mse_error)
 
-        train_mae_error_list.append(train_mae_error)
         validation_mae_error_list.append(val_mae_error)
 
-        train_corr_list.append(train_corr[0])
         validation_corr_list.append(val_corr[0])
 
-        train_corr_pval_list.append(train_corr[1])
         validation_corr_pval_list.append(val_corr[1])
 
     return report_results(
-        train_mse_error_list,
         validation_mse_error_list,
-        train_mae_error_list,
         validation_mae_error_list,
-        train_corr_list,
         validation_corr_list,
-        train_corr_pval_list,
         validation_corr_pval_list,
     )