feat: optimize Gradient Boosting and XGBoost parameters for improved model performance

app/training/train_classifier.py

@@ -265,10 +265,12 @@ def _build_candidates() -> list[tuple[str, Any]]:
         (
             "Gradient Boosting",
             GradientBoostingClassifier(
-                n_estimators=200,
-                max_depth=6,
+                n_estimators=150,
+                max_depth=3,
                 learning_rate=0.1,
-                subsample=0.8,
+                subsample=0.7,
+                min_samples_leaf=20,
+                min_samples_split=30,
                 random_state=42,
             ),
         ),
@@ -303,15 +305,15 @@ def _build_candidates() -> list[tuple[str, Any]]:
         candidates.append((
             "XGBoost",
             xgb.XGBClassifier(
-                n_estimators=200,
-                max_depth=5,
-                learning_rate=0.05,
-                subsample=0.8,
-                colsample_bytree=0.8,
-                min_child_weight=5,
-                reg_alpha=0.1,
-                reg_lambda=1.0,
-                gamma=0.1,
+                n_estimators=150,
+                max_depth=3,
+                learning_rate=0.1,
+                subsample=0.7,
+                colsample_bytree=0.7,
+                min_child_weight=10,
+                reg_alpha=0.5,
+                reg_lambda=2.0,
+                gamma=0.3,
                 scale_pos_weight=1.0,
                 eval_metric="logloss",
                 random_state=42,
@@ -323,15 +325,15 @@ def _build_candidates() -> list[tuple[str, Any]]:
         candidates.append((
             "LightGBM",
             lgb.LGBMClassifier(
-                n_estimators=200,
-                max_depth=5,
-                learning_rate=0.05,
-                num_leaves=24,
-                subsample=0.8,
-                colsample_bytree=0.8,
-                min_child_weight=5,
-                reg_alpha=0.1,
-                reg_lambda=1.0,
+                n_estimators=150,
+                max_depth=3,
+                learning_rate=0.1,
+                num_leaves=8,
+                subsample=0.7,
+                colsample_bytree=0.7,
+                min_child_weight=10,
+                reg_alpha=0.5,
+                reg_lambda=2.0,
                 class_weight="balanced",
                 random_state=42,
                 verbose=-1,