مشروع: التنبؤ بأسعار المنازل

import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor
from sklearn.linear_model import Ridge
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.pipeline import Pipeline
from sklearn.impute import SimpleImputer
import joblib
import warnings
warnings.filterwarnings("ignore")

print("=" * 60)
print("🏠 مشروع: التنبؤ بأسعار المنازل — Pipeline كامل")
print("=" * 60)

# ─────────────────────────────────────────
# 1. توليد بيانات واقعية
# ─────────────────────────────────────────
np.random.seed(42)
n = 2000

districts = ["الرياض - العليا", "الرياض - النرجس", "الرياض - حطين",
             "الرياض - الملقا", "الرياض - الروضة"]

data = pd.DataFrame({
    "area_sqm":     np.random.normal(200, 80, n).clip(60, 600),
    "rooms":        np.random.randint(2, 8, n),
    "bathrooms":    np.random.randint(1, 5, n),
    "floors":       np.random.randint(1, 4, n),
    "age_years":    np.random.randint(0, 30, n),
    "has_pool":     np.random.binomial(1, 0.2, n),
    "has_garden":   np.random.binomial(1, 0.4, n),
    "parking":      np.random.randint(0, 4, n),
    "district":     np.random.choice(districts, n),
    "dist_center_km": np.random.exponential(8, n).clip(1, 30),
})

# قيم مفقودة (واقعية)
data.loc[np.random.choice(n, 50, replace=False), "bathrooms"] = np.nan
data.loc[np.random.choice(n, 30, replace=False), "age_years"]  = np.nan

# السعر الحقيقي
district_premium = {d: p for d, p in zip(districts, [1.3, 1.1, 1.2, 0.9, 1.0])}
data["price"] = (
    data["area_sqm"] * 3500
    + data["rooms"] * 25000
    + data["bathrooms"].fillna(2) * 15000
    - data["age_years"].fillna(10) * 2000
    + data["has_pool"] * 80000
    + data["has_garden"] * 40000
    + data["parking"] * 20000
    - data["dist_center_km"] * 5000
    + data["district"].map(district_premium) * 50000
    + np.random.normal(0, 50000, n)
).clip(200000, 5000000)

print(f"\n1. البيانات:")
print(f"   عدد المنازل     : {len(data)}")
print(f"   الميزات          : {data.shape[1] - 1}")
print(f"   قيم مفقودة      : {data.isnull().sum().sum()}")
print(f"   نطاق الأسعار    : {data['price'].min():,.0f} - {data['price'].max():,.0f} ريال")
print(f"   متوسط السعر     : {data['price'].mean():,.0f} ريال")

# ─────────────────────────────────────────
# 2. Feature Engineering
# ─────────────────────────────────────────
print("\n2. Feature Engineering:")
data["price_per_sqm"] = data["price"] / data["area_sqm"]  # للتحليل فقط
data["total_amenities"] = data["has_pool"] + data["has_garden"] + (data["parking"] > 0)
data["rooms_per_floor"] = data["rooms"] / data["floors"]

le = LabelEncoder()
data["district_encoded"] = le.fit_transform(data["district"])

features = ["area_sqm", "rooms", "bathrooms", "floors", "age_years",
            "has_pool", "has_garden", "parking", "dist_center_km",
            "district_encoded", "total_amenities", "rooms_per_floor"]

X = data[features]
y = data["price"]

print(f"   ميزات بعد الهندسة: {len(features)}")

# ─────────────────────────────────────────
# 3. مقارنة النماذج
# ─────────────────────────────────────────
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

models = {
    "Ridge Regression": Pipeline([
        ("imputer", SimpleImputer(strategy="median")),
        ("scaler", StandardScaler()),
        ("model", Ridge(alpha=100))
    ]),
    "Random Forest": Pipeline([
        ("imputer", SimpleImputer(strategy="median")),
        ("model", RandomForestRegressor(n_estimators=200, n_jobs=-1, random_state=42))
    ]),
    "Gradient Boosting": Pipeline([
        ("imputer", SimpleImputer(strategy="median")),
        ("model", GradientBoostingRegressor(n_estimators=200, random_state=42))
    ]),
}

print("\n3. مقارنة النماذج (CV=5):")
print(f"  {'النموذج':25s} | {'MAE':>12s} | {'R²':>8s}")
print("  " + "-" * 55)

best_model_name, best_r2 = None, -999
for name, pipe in models.items():
    r2_scores  = cross_val_score(pipe, X_train, y_train, cv=5, scoring="r2", n_jobs=-1)
    mae_scores = cross_val_score(pipe, X_train, y_train, cv=5,
                                 scoring="neg_mean_absolute_error", n_jobs=-1)
    r2  = r2_scores.mean()
    mae = -mae_scores.mean()
    print(f"  {name:25s} | {mae:>12,.0f} | {r2:>8.4f}")
    if r2 > best_r2:
        best_r2, best_model_name = r2, name

print(f"\n  ✅ أفضل نموذج: {best_model_name}")

# ─────────────────────────────────────────
# 4. التدريب النهائي والتقييم
# ─────────────────────────────────────────
final_model = models[best_model_name]
final_model.fit(X_train, y_train)
y_pred = final_model.predict(X_test)

mae  = mean_absolute_error(y_test, y_pred)
rmse = np.sqrt(mean_squared_error(y_test, y_pred))
r2   = r2_score(y_test, y_pred)

print("\n4. نتائج Test Set النهائية:")
print(f"   MAE  = {mae:>10,.0f} ريال")
print(f"   RMSE = {rmse:>10,.0f} ريال")
print(f"   R²   = {r2:>10.4f}")
print(f"   الخطأ النسبي = {mae/y_test.mean()*100:.1f}%")

# ─────────────────────────────────────────
# 5. حفظ النموذج
# ─────────────────────────────────────────
joblib.dump(final_model, "house_price_model.pkl")
print("\n5. ✅ تم حفظ النموذج في: house_price_model.pkl")

# ─────────────────────────────────────────
# 6. تنبؤ ببيت جديد
# ─────────────────────────────────────────
loaded_model = joblib.load("house_price_model.pkl")

new_house = pd.DataFrame([{
    "area_sqm": 280, "rooms": 5, "bathrooms": 3, "floors": 2,
    "age_years": 5, "has_pool": 1, "has_garden": 1, "parking": 2,
    "dist_center_km": 8, "district_encoded": 2,
    "total_amenities": 2, "rooms_per_floor": 2.5
}])

predicted_price = loaded_model.predict(new_house)[0]
print(f"\n6. تنبؤ لبيت جديد:")
print(f"   الخصائص: 280م², 5 غرف, مسبح, حديقة, 8كم من المركز")
print(f"   السعر المتوقع: {predicted_price:,.0f} ريال")
print("\n🎉 مبروك! أكملت دورة تعلم الآلة بنجاح")