stacking模型融合代码

好的，下面是一个 stacking 模型融合的示例代码：

首先，我们需要导入必要的库和数据集：

import pandas as pd
import numpy as np
from sklearn.model_selection import KFold
from sklearn.metrics import mean_squared_error
from sklearn.ensemble import RandomForestRegressor, GradientBoostingRegressor
from xgboost import XGBRegressor

# 读取训练数据和测试数据
train_df = pd.read_csv('train.csv')
test_df = pd.read_csv('test.csv')

然后，我们定义一个函数来进行交叉验证，训练模型和预测：

def run_model(model, train_X, train_y, test_X):
    model.fit(train_X, train_y)
    y_pred = model.predict(test_X)
    return y_pred

接下来，我们定义一个 stacking 模型融合的函数：

def stacking(models, train_X, train_y, test_X, n_fold):
    # 预测结果矩阵
    train_pred = np.zeros((train_X.shape[0], len(models)))
    test_pred = np.zeros((test_X.shape[0], len(models)))
    kf = KFold(n_splits=n_fold, shuffle=True, random_state=42)

    for i, model in enumerate(models):
        # 预测测试数据
        test_pred_i = np.zeros((test_X.shape[0], n_fold))
        for j, (train_idx, val_idx) in enumerate(kf.split(train_X)):
            # 划分训练集和验证集
            train_X_fold, train_y_fold = train_X[train_idx], train_y[train_idx]
            val_X_fold, val_y_fold = train_X[val_idx], train_y[val_idx]
            # 训练模型并预测验证集和测试集
            y_val_pred_fold = run_model(model, train_X_fold, train_y_fold, val_X_fold)
            y_test_pred_fold = run_model(model, train_X_fold, train_y_fold, test_X)
            # 记录预测结果
            train_pred[val_idx, i] = y_val_pred_fold
            test_pred_i[:, j] = y_test_pred_fold
        # 对测试集的预测结果取平均
        test_pred[:, i] = test_pred_i.mean(axis=1)
    
    # 使用最终的模型对测试集进行预测
    final_model = XGBRegressor()
    final_model.fit(train_pred, train_y)
    y_pred = final_model.predict(test_pred)
    return y_pred

最后，我们定义一些模型并调用 stacking 函数进行模型融合：

# 定义模型
rf_model = RandomForestRegressor(random_state=42)
gbdt_model = GradientBoostingRegressor(random_state=42)
xgb_model = XGBRegressor(random_state=42)

# 进行 stacking 模型融合
models = [rf_model, gbdt_model, xgb_model]
n_fold = 5
y_pred = stacking(models, train_X, train_y, test_X, n_fold)

这就是一个简单的 stacking 模型融合的示例代码，你可以根据自己的数据和模型进行修改。