Getting cv metric gen ready

model_selection/cv_metric_gen.py

+"""
+    Metric generation for each tuned model.
+    Done in a different script for perfomance and clarity purposes.
+"""
+
+# Libraries
+# --------------------------------------------------------------------------------------------------------
+import pandas as pd
+import numpy as np
+from xgboost import XGBClassifier
+from sklearn.metrics import confusion_matrix
+from sklearn.metrics import f1_score, make_scorer, precision_score, recall_score, accuracy_score, roc_auc_score, average_precision_score
+from sklearn.ensemble import RandomForestClassifier, BaggingClassifier, AdaBoostClassifier
+from sklearn.neural_network import MLPClassifier
+from sklearn.svm import SVC
+from sklearn.linear_model import  LogisticRegression
+from sklearn.tree import DecisionTreeClassifier
+from sklearn.model_selection import StratifiedKFold, cross_validate
+from sklearn.metrics import RocCurveDisplay, roc_curve
+from sklearn.metrics import PrecisionRecallDisplay, precision_recall_curve
+import matplotlib.pyplot as plt
+import ast # String to dictionary
+# --------------------------------------------------------------------------------------------------------
+
+# Function to read training datasets
+# --------------------------------------------------------------------------------------------------------
+def read_data():
+
+    # Load ORIGINAL training data
+    X_train_pre = np.load('../gen_train_data/data/output/pre/X_train_pre.npy', allow_pickle=True)
+    y_train_pre = np.load('../gen_train_data/data/output/pre/y_train_pre.npy', allow_pickle=True)
+    X_train_post = np.load('../gen_train_data/data/output/post/X_train_post.npy', allow_pickle=True)
+    y_train_post = np.load('../gen_train_data/data/output/post/y_train_post.npy', allow_pickle=True)
+
+    # Load oversampled training data
+    X_train_over_pre = np.load('../gen_train_data/data/output/pre/X_train_over_pre.npy', allow_pickle=True)
+    y_train_over_pre = np.load('../gen_train_data/data/output/pre/y_train_over_pre.npy', allow_pickle=True)
+    X_train_over_post = np.load('../gen_train_data/data/output/post/X_train_over_post.npy', allow_pickle=True)
+    y_train_over_post = np.load('../gen_train_data/data/output/post/y_train_over_post.npy', allow_pickle=True)
+
+    # Load undersampled training data
+    X_train_under_pre = np.load('../gen_train_data/data/output/pre/X_train_under_pre.npy', allow_pickle=True)
+    y_train_under_pre = np.load('../gen_train_data/data/output/pre/y_train_under_pre.npy', allow_pickle=True)
+    X_train_under_post = np.load('../gen_train_data/data/output/post/X_train_under_post.npy', allow_pickle=True)
+    y_train_under_post = np.load('../gen_train_data/data/output/post/y_train_under_post.npy', allow_pickle=True)
+
+    data_dic = {
+        "X_train_pre": X_train_pre,
+        "y_train_pre": y_train_pre,
+        "X_train_post": X_train_post,
+        "y_train_post": y_train_post,
+        "X_train_over_pre": X_train_over_pre,
+        "y_train_over_pre": y_train_over_pre,
+        "X_train_over_post": X_train_over_post,
+        "y_train_over_post": y_train_over_post,
+        "X_train_under_pre": X_train_under_pre,
+        "y_train_under_pre": y_train_under_pre,
+        "X_train_under_post": X_train_under_post,
+        "y_train_under_post": y_train_under_post,
+    }
+
+    return data_dic
+# --------------------------------------------------------------------------------------------------------
+
+# Returning tuned models for each situation
+# --------------------------------------------------------------------------------------------------------
+def get_tuned_models(group_str, method_str):
+    # Read sheet corresponding to group and method with tuned models and their hyperparam
+    tuned_models_df = pd.read_excel("./output_hyperparam/hyperparamers.xlsx",sheet_name=f"{group_str}_{method_str}")
+    # Mapping from model abbreviations to sklearn model classes
+    model_mapping = {
+        'DT': DecisionTreeClassifier,
+        'RF': RandomForestClassifier,
+        'Bagging': BaggingClassifier,
+        'AB': AdaBoostClassifier,
+        'XGB': XGBClassifier,
+        'LR': LogisticRegression,
+        'SVM': SVC,
+        'MLP': MLPClassifier
+    }
+    tuned_models = {}
+    # Iterate through each row of the DataFrame
+    for _, row in tuned_models_df.iterrows():
+        model_name = row[0]
+        # Read dictionary
+        parameters = ast.literal_eval(row['Parameters'])
+        # Add extra parameters 
+        if model_name == 'AB':
+            parameters['algorithm'] = 'SAMME'
+        elif model_name == 'LR':
+            parameters['max_iter'] = 1000
+        elif model_name == 'SVM':
+            parameters['max_iter'] = 1000
+            parameters['probability'] = True
+        elif model_name == "MLP":
+            parameters['max_iter'] = 500
+        # Add class_weight argument for cost-sensitive learning method
+        if 'CW' in method_str:
+            if model_name == 'Bagging' or model_name == 'AB':
+                parameters['estimator'] = DecisionTreeClassifier(class_weight='balanced')
+            else:
+                parameters['class_weight'] = 'balanced'
+        # Fetch class
+        model_class = model_mapping[model_name]
+        # Initialize model
+        tuned_models[model_name] = model_class(**parameters)
+    return tuned_models
+# --------------------------------------------------------------------------------------------------------
+
+# Scorers
+# --------------------------------------------------------------------------------------------------------
+def TN_scorer(clf, X, y):
+    """Gives the number of samples predicted as true negatives"""
+    y_pred = clf.predict(X)
+    cm = confusion_matrix(y, y_pred)
+    TN = cm[0,0]
+    return TN
+def FN_scorer(clf, X, y):
+    """Gives the number of samples predicted as false negatives"""
+    y_pred = clf.predict(X)
+    cm = confusion_matrix(y, y_pred)
+    FN = cm[0,1]
+    return FN
+def FP_scorer(clf, X, y):
+    """Gives the number of samples predicted as false positive"""
+    y_pred = clf.predict(X)
+    cm = confusion_matrix(y, y_pred)
+    FP = cm[1,0]
+    return FP
+def TP_scorer(clf, X, y):
+    """Gives the number of samples predicted as true positive"""
+    y_pred = clf.predict(X)
+    cm = confusion_matrix(y, y_pred)
+    TP = cm[1,1]
+    return TP
+
+def negative_recall_scorer(clf, X, y):
+    """Gives the negative recall defined as the (number of true_negative_samples)/(total number of negative samples)"""
+    y_pred = clf.predict(X)
+    cm = confusion_matrix(y, y_pred)
+    TN_prop = cm[0,0]/(cm[0,1]+cm[0,0])
+    return TN_prop
+# --------------------------------------------------------------------------------------------------------
+
+if __name__ == "__main__":
+
+    # Setup
+    # --------------------------------------------------------------------------------------------------------
+    # Reading training data
+    data_dic = read_data()
+    # Scorings to use for cv metric generation
+    scorings = {
+        'F1':make_scorer(f1_score), 
+        'PREC':make_scorer(precision_score), 
+        'REC':make_scorer(recall_score), 
+        # 'ACC': make_scorer(accuracy_score),
+        # 'NREC': negative_recall_scorer, 
+        # 'TN':TN_scorer, 
+        # 'FN':FN_scorer, 
+        # 'FP':FP_scorer, 
+        # 'TP':TP_scorer,
+        # 'AUROC': make_scorer(roc_auc_score), 
+        # 'AUPRC': make_scorer(average_precision_score)
+        } 
+    method_names = {
+        0: "ORIG",
+        1: "ORIG_CW",
+        2: "OVER",
+        3: "UNDER"
+    }
+    # Defining cross-validation protocol
+    cv = StratifiedKFold(n_splits=10, shuffle=True, random_state=1) 
+    # --------------------------------------------------------------------------------------------------------
+
+    # Metric generation through cv for tuned models3
+    # --------------------------------------------------------------------------------------------------------
+    scores_sheets = {} # To store score dfs as sheets in the same excel file
+    for i, group in enumerate(['pre', 'post']):
+        for j, method in enumerate(['', '', 'over_', 'under_']):
+            print(f"{group}-{method_names[j]}")
+            # Get train dataset based on group and method
+            X_train = data_dic['X_train_' + method + group]
+            y_train = data_dic['y_train_' + method + group]
+            # Get tuned models for this group and method
+            models = get_tuned_models(group, method_names[j])
+            # Scores df
+            # One column per cv split, one row for each model-metric
+            scores_df = pd.DataFrame(columns=range(1,11), index=[f"{model_name}_{metric_name}" for model_name in models.keys() for metric_name in scorings.keys()])
+            # Metric generation for each model
+            for model_name, model in models.items():
+                # Retrieve cv scores for our metrics of interest
+                scores = cross_validate(model, X_train, y_train, scoring=scorings, cv=cv, return_train_score=True, n_jobs=10)
+                # Save results of each fold
+                scores_df.loc[model_name +'_F1']=list(np.around(np.array(scores["test_f1"]),4)) 
+                scores_df.loc[model_name +'_PREC']=list(np.around(np.array(scores["test_precision"]),4))
+                scores_df.loc[model_name +'_REC']=list(np.around(np.array(scores["test_recall"]),4))
+            # Store the DataFrame in the dictionary with a unique key for each sheet
+            sheet_name = f"{group}_{method_names[j]}"
+            scores_sheets[sheet_name] = scores_df
+    # Write results to Excel file
+    with pd.ExcelWriter('./output_cv_metrics.xlsx') as writer:
+        for sheet_name, data in scores_sheets.items():
+            data.to_excel(writer, sheet_name=sheet_name)
+    print("Successful metric generation for tuned models")
+
+                
+            
\ No newline at end of file