مشروع: توقع أسعار المنازل

1. إنشاء البيانات
2. استكشاف سريع
3. فصل X (مع ميزات رقمية + فئوية) و y
4. بناء Pipeline مع ColumnTransformer
5. تقسيم train/test
6. تدريب LinearRegression
7. تنبّؤ + تقييم (MAE, RMSE, R²)
8. تفسير المعاملات

import pandas as pd

data = {
    'area':    [80, 120, 150, 100, 200, 90, 250, 180, 110, 300, 140, 220],
    'rooms':   [2, 3, 4, 2, 4, 2, 5, 4, 3, 5, 3, 4],
    'age':     [20, 15, 10, 25, 5, 30, 3, 12, 18, 2, 22, 8],
    'neighborhood': ['A', 'B', 'A', 'C', 'B', 'A', 'B', 'C', 'A', 'B', 'C', 'A'],
    'price':   [200, 320, 380, 240, 480, 220, 580, 420, 280, 720, 350, 510]
}
df = pd.DataFrame(data)

print(f"عدد المنازل: {len(df)}")
print(df.head())

print(df.describe())
print()
print("عدد القيم المفقودة:")
print(df.isna().sum())
print()
print("الأسعار حسب الحي:")
print(df.groupby('neighborhood')['price'].mean())

# الميزات (4): 3 رقمية + 1 فئوية
X = df.drop('price', axis=1)

# الهدف
y = df['price']

print(f"شكل X: {X.shape}")
print(f"شكل y: {y.shape}")

from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.linear_model import LinearRegression

# Pipeline preprocessing
preprocessor = ColumnTransformer(
    transformers=[
        ('num', StandardScaler(), ['area', 'rooms', 'age']),
        ('cat', OneHotEncoder(handle_unknown='ignore'), ['neighborhood'])
    ]
)

# Pipeline كامل
pipe = Pipeline([
    ('preprocess', preprocessor),
    ('model', LinearRegression())
])

from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.25, random_state=42
)
print(f"Train: {len(X_train)} منزل")
print(f"Test:  {len(X_test)} منزل")

pipe.fit(X_train, y_train)

from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
import numpy as np

# تنبّؤ
y_pred = pipe.predict(X_test)

# تقييم
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(f"MAE:  {mae:.2f} (آلاف)")
print(f"RMSE: {rmse:.2f} (آلاف)")
print(f"R²:   {r2:.3f}")

print("\nمقارنة (فعلي vs تنبّؤ):")
for actual, predicted in zip(y_test, y_pred):
    print(f"  {actual:5.0f}  vs  {predicted:6.1f}")

# الوصول إلى LinearRegression داخل الـ Pipeline
model = pipe.named_steps['model']
preprocessor = pipe.named_steps['preprocess']

# أسماء الميزات بعد OneHot
feature_names_num = ['area', 'rooms', 'age']
feature_names_cat = list(preprocessor.named_transformers_['cat'].get_feature_names_out(['neighborhood']))
all_feature_names = feature_names_num + feature_names_cat

print("معاملات النموذج:")
for name, coef in zip(all_feature_names, model.coef_):
    direction = "يزيد السعر" if coef > 0 else "يقلل السعر"
    print(f"  {name:15s}: {coef:+8.2f}  ({direction})")
print(f"  {'(intercept)':15s}: {model.intercept_:+.2f}")

from sklearn.linear_model import Ridge
pipe_ridge = Pipeline([
    ('preprocess', preprocessor),
    ('model', Ridge(alpha=1.0))
])
pipe_ridge.fit(X_train, y_train)
print(f"Ridge R²: {pipe_ridge.score(X_test, y_test):.3f}")

from sklearn.ensemble import RandomForestRegressor
pipe_rf = Pipeline([
    ('preprocess', preprocessor),
    ('model', RandomForestRegressor(n_estimators=100, max_depth=3, random_state=42))
])
pipe_rf.fit(X_train, y_train)
print(f"RF R²: {pipe_rf.score(X_test, y_test):.3f}")

from sklearn.model_selection import cross_val_score

scores = cross_val_score(pipe, X, y, cv=5, scoring='r2')
print(f"CV R²: {scores.mean():.3f} ± {scores.std():.3f}")