Update train_hybrid_model.py

Hybrid_model_code/train_hybrid_model.py

@@ -5,57 +5,46 @@ from sklearn.feature_extraction.text import TfidfVectorizer
 from sklearn.model_selection import train_test_split
 from xgboost import XGBClassifier
 from sklearn.metrics import accuracy_score, classification_report
-from scipy.sparse import hstack, csr_matrix # لدمج مصفوفات البيانات المتفرقة
+from scipy.sparse import hstack, csr_matrix
 
-# --- 1. الإعدادات ---
-# ملف الخصائص اللغوية (الناتج من Farasa)
 LINGUISTIC_FILE = "features_gemini_vs_human_augmented.csv"
-# ملف النصوص الخام (لاستخراج N-Grams)
 RAW_DATA_FILE = "merged_dataset_clean2.csv" 
 COL_HUMAN = "human_collected_dataset"
 COL_GEMINI = "gemini_rephrased_v2_5"
 
-print("🚀 Starting Hybrid Model Training (Farasa + N-Grams)...")
+print("Starting hybrid model training (Farasa + N-grams)...")
 
-# --- 2. تحميل وتجهيز البيانات ---
-
-# أ. تحميل الخصائص اللغوية
 if not os.path.exists(LINGUISTIC_FILE):
-    print(f"❌ Error: Linguistic features file '{LINGUISTIC_FILE}' not found. Run previous pipeline first.")
+    print(f"Error: linguistic features file '{LINGUISTIC_FILE}' not found. Run the previous pipeline first.")
     exit()
 
 df_features = pd.read_csv(LINGUISTIC_FILE)
 df_features.dropna(inplace=True)
 X_linguistic = df_features.drop(columns=['label'])
 y = df_features['label']
-print(f"✅ Loaded Linguistic Features: {len(X_linguistic)} samples.")
-
+print(f"Loaded linguistic features: {len(X_linguistic)} samples.")
 
-# ب. تحميل النصوص الخام (لـ N-Grams)
 try:
     df_raw = pd.read_csv(RAW_DATA_FILE)
     df_raw.columns = df_raw.columns.str.strip()
     
-    # دمج النصوص الخام بنفس الترتيب
     df_human = pd.DataFrame({'text': df_raw[COL_HUMAN]})
     df_ai = pd.DataFrame({'text': df_raw[COL_GEMINI]})
     df_text = pd.concat([df_human, df_ai], ignore_index=True)
     
-    # يجب أن تتطابق الأحجام
     min_len = min(len(df_features), len(df_text))
     df_text = df_text.iloc[:min_len]
     X_linguistic = X_linguistic.iloc[:min_len]
     y = y.iloc[:min_len]
     
     X_text = df_text['text'].astype(str)
-    print(f"✅ Loaded Raw Text: {len(X_text)} samples (Synced).")
+    print(f"Loaded raw text: {len(X_text)} samples (synced).")
 
 except Exception as e:
-    print(f"❌ Error loading raw text data: {e}")
+    print(f"Error loading raw text data: {e}")
     exit()
 
-# 3. استخراج خصائص N-Grams
-print("⚙️  Generating N-Gram Features (TF-IDF Character N-grams)...")
+print("Generating N-gram features (TF-IDF character n-grams)...")
 
 tfidf = TfidfVectorizer(
     analyzer='char',
@@ -64,24 +53,19 @@ tfidf = TfidfVectorizer(
     min_df=5               
 )
 X_ngrams = tfidf.fit_transform(X_text)
-print(f"✅ N-Grams Features Shape: {X_ngrams.shape}")
+print(f"N-gram features shape: {X_ngrams.shape}")
 
-# 4. دمج الخصائص (Hybrid Concatenation)
-# نحول خصائص فراسة إلى مصفوفة متفرقة (Sparse Matrix) لدمجها مع TF-IDF
 X_linguistic_sparse = csr_matrix(X_linguistic.values)
 
-# الدمج الأفقي (الخصائص اللغوية + خصائص N-Grams)
 X_hybrid = hstack([X_ngrams, X_linguistic_sparse])
 
-print(f"✅ Hybrid Dataset Shape: {X_hybrid.shape}")
+print(f"Hybrid dataset shape: {X_hybrid.shape}")
 
-# 5. تقسيم البيانات
 X_train, X_test, y_train, y_test = train_test_split(
     X_hybrid, y, test_size=0.2, random_state=42, stratify=y
 )
 
-# 6. تدريب XGBoost (باستخدام أفضل المعلمات التي وجدتها سابقاً)
-print("🤖 Training XGBoost Hybrid Model...")
+print("Training XGBoost hybrid model...")
 
 model = XGBClassifier(
     n_estimators=500, 
@@ -93,13 +77,12 @@ model = XGBClassifier(
 )
 model.fit(X_train, y_train)
 
-# 7. التقييم
 y_pred = model.predict(X_test)
 accuracy = accuracy_score(y_test, y_pred)
 
 print("\n" + "="*50)
-print(f"🏆 Final Hybrid Model Accuracy: {accuracy * 100:.2f}%")
+print(f"Final hybrid model accuracy: {accuracy * 100:.2f}%")
 print("="*50)
 
-print("\n📋 Classification Report:")
-print(classification_report(y_test, y_pred, target_names=['Human', 'AI']))
+print("\nClassification report:")
+print(classification_report(y_test, y_pred, target_names=['Human', 'AI']))