ai-data-analysis-guided-tour

import pandas as pd

df = pd.read_csv("/Users/wyatt/Downloads/Housing.csv")

print(f"Shape: {df.shape}")
columns = df.columns.tolist()
print(f"Columns ({len(columns)}):", columns if len(columns) <= 100 else columns[:100] + ["..."])

df.head()

quality = pd.DataFrame({
    "dtype": df.dtypes.astype(str),
    "missing": df.isna().sum(),
    "unique_values": df.nunique()
})
quality

from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import OneHotEncoder, StandardScaler
from sklearn.linear_model import LinearRegression, Ridge
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor, ExtraTreesRegressor
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score
import pandas as pd

X = df.drop(columns="price")
y = df["price"]

numeric_features = X.select_dtypes(include="number").columns.tolist()
categorical_features = X.select_dtypes(exclude="number").columns.tolist()

preprocessor = ColumnTransformer([
    ("numeric", StandardScaler(), numeric_features),
    ("categorical", OneHotEncoder(handle_unknown="ignore", sparse_output=False), categorical_features)
])

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42
)

models = {
    "线性回归": LinearRegression(),
    "岭回归": Ridge(alpha=10),
    "随机森林": RandomForestRegressor(n_estimators=400, random_state=42, n_jobs=-1),
    "梯度提升树": GradientBoostingRegressor(random_state=42),
    "极端随机树": ExtraTreesRegressor(n_estimators=400, random_state=42, n_jobs=-1)
}

fitted_models = {}
results = []

for name, model in models.items():
    pipeline = Pipeline([("preprocess", preprocessor), ("model", model)])
    pipeline.fit(X_train, y_train)
    prediction = pipeline.predict(X_test)
    
    cv_rmse = -cross_val_score(
        pipeline, X, y, cv=5, scoring="neg_root_mean_squared_error"
    ).mean()
    
    fitted_models[name] = pipeline
    results.append({
        "模型": name,
        "测试集 R²": r2_score(y_test, prediction),
        "测试集 MAE": mean_absolute_error(y_test, prediction),
        "测试集 RMSE": mean_squared_error(y_test, prediction) ** 0.5,
        "5折交叉验证 RMSE": cv_rmse
    })

results_df = pd.DataFrame(results).sort_values("测试集 RMSE").reset_index(drop=True)
results_df.round(2)

from xgboost import XGBRegressor

xgb_model = XGBRegressor(
    objective="reg:squarederror",
    n_estimators=500,
    learning_rate=0.03,
    max_depth=3,
    subsample=0.8,
    colsample_bytree=0.8,
    reg_lambda=1.0,
    random_state=42,
    n_jobs=-1
)

xgb_pipeline = Pipeline([
    ("preprocess", preprocessor),
    ("model", xgb_model)
])

xgb_pipeline.fit(X_train, y_train)
xgb_prediction = xgb_pipeline.predict(X_test)

xgb_cv_rmse = -cross_val_score(
    xgb_pipeline, X, y, cv=5, scoring="neg_root_mean_squared_error"
).mean()

fitted_models["XGBoost"] = xgb_pipeline

xgb_result = pd.DataFrame([{
    "模型": "XGBoost",
    "测试集 R²": r2_score(y_test, xgb_prediction),
    "测试集 MAE": mean_absolute_error(y_test, xgb_prediction),
    "测试集 RMSE": mean_squared_error(y_test, xgb_prediction) ** 0.5,
    "5折交叉验证 RMSE": xgb_cv_rmse
}])

comparison_df = (
    pd.concat([results_df, xgb_result], ignore_index=True)
    .sort_values("测试集 RMSE")
    .reset_index(drop=True)
)

comparison_df.round(2)

xgb_pipeline.fit(X_train, y_train)
xgb_prediction = xgb_pipeline.predict(X_test)

xgb_result = pd.DataFrame([{
    "模型": "XGBoost",
    "测试集 R²": r2_score(y_test, xgb_prediction),
    "测试集 MAE": mean_absolute_error(y_test, xgb_prediction),
    "测试集 RMSE": mean_squared_error(y_test, xgb_prediction) ** 0.5,
    "5折交叉验证 RMSE": -cross_val_score(
        xgb_pipeline, X, y, cv=5, scoring="neg_root_mean_squared_error"
    ).mean()
}])

pd.concat([results_df, xgb_result], ignore_index=True).sort_values(
    "测试集 RMSE"
).reset_index(drop=True).round(2)