# Encode labels for models that require it
le = LabelEncoder()
y_train_encoded = le.fit_transform(y_train)
y_test_encoded = le.transform(y_test)
best_models = {}
tuned_model_results = {}
# Logistic Regression
start_time = time.time()
lr = LogisticRegression(max_iter=2000, random_state=42, n_jobs=1)
param_grid_lr = {'C': [0.01, 0.05, 0.1, 0.5]}
grid_lr = GridSearchCV(lr, param_grid_lr, cv=3, n_jobs=32, scoring='accuracy')
grid_lr.fit(X_train_scaled, y_train_encoded)
best_models['Logistic Regression'] = grid_lr.best_estimator_
print(f"Finished in {time.time() - start_time:.2f}s. Best Score: {grid_lr.best_score_:.4f}")
print(f"Best Parameters: {grid_lr.best_params_}")
# K-Nearest Neighbors
start_time = time.time()
knn = KNeighborsClassifier()
param_grid_knn = {'n_neighbors': [3, 5, 7]}
grid_knn = GridSearchCV(knn, param_grid_knn, cv=3, n_jobs=32, scoring='accuracy')
grid_knn.fit(X_train_scaled, y_train_encoded)
best_models['K-Nearest Neighbors'] = grid_knn.best_estimator_
print(f"Finished in {time.time() - start_time:.2f}s. Best Score: {grid_knn.best_score_:.4f}")
print(f"Best Parameters: {grid_knn.best_params_}")
# Random Forest
start_time = time.time()
rf = RandomForestClassifier(random_state=42, n_jobs=1)
param_grid_rf = {'n_estimators': [100, 200], 'max_depth': [10, 20, None]}
grid_rf = GridSearchCV(rf, param_grid_rf, cv=3, n_jobs=32, scoring='accuracy')
grid_rf.fit(X_train_scaled, y_train_encoded)
best_models['Random Forest'] = grid_rf.best_estimator_
print(f"Finished in {time.time() - start_time:.2f}s. Best Score: {grid_rf.best_score_:.4f}")
print(f"Best Parameters: {grid_rf.best_params_}")
start_time = time.time()
warnings.filterwarnings("ignore", category=UserWarning, module='xgboost')
param_grid_xgb = {'learning_rate': [0.05, 0.1], 'max_depth': [5, 10], 'n_estimators': [100, 200]}
xgb = XGBClassifier(device="cuda", random_state=42, use_label_encoder=False, eval_metric='mlogloss')
grid_xgb = GridSearchCV(xgb, param_grid_xgb, cv=3, n_jobs=1, scoring='accuracy')
grid_xgb.fit(X_train_scaled, y_train_encoded)
best_models['XGBoost'] = grid_xgb.best_estimator_
print(f"Finished in {time.time() - start_time:.2f}s. Best Score: {grid_xgb.best_score_:.4f}")
print(f"Best Parameters: {grid_xgb.best_params_}")
# LinearSVC
start_time = time.time()
param_grid_lsvc = {'C': [0.01, 0.1, 1.0]}
lsvc = LinearSVC(max_iter=4000, random_state=42, dual='auto')
grid_lsvc = GridSearchCV(lsvc, param_grid_lsvc, cv=3, n_jobs=32, scoring='accuracy')
grid_lsvc.fit(X_train_scaled, y_train_encoded)
best_models['LinearSVC'] = grid_lsvc.best_estimator_
print(f"Finished in {time.time() - start_time:.2f}s. Best Score: {grid_lsvc.best_score_:.4f}")
print(f"Best Parameters: {grid_lsvc.best_params_}")
