Sklearn -发现样本数量不一致的输入变量：[16512,4128]

import pandas as pd

# 假设你有两个DataFrame
df1 = pd.read_csv('data1.csv')
df2 = pd.read_csv('data2.csv')

# 检查样本数量
print(len(df1), len(df2))

# 假设df1和df2有相同的列名，只是样本数量不同
merged_df = pd.concat([df1, df2], axis=1)

# 对齐样本
aligned_df1, aligned_df2 = df1.align(df2, join='inner')

# 示例：确保在特征提取过程中没有丢失样本
def extract_features(data):
    # 假设这是一个特征提取函数
    features = data.dropna()  # 确保没有丢失样本
    return features

features1 = extract_features(df1)
features2 = extract_features(df2)

# 检查样本数量
print(len(features1), len(features2))

from sklearn.compose import ColumnTransformer
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler

# 假设你有两个特征矩阵X1和X2
preprocessor = ColumnTransformer(
    transformers=[
        ('num1', StandardScaler(), X1),
        ('num2', StandardScaler(), X2)
    ])

# 确保X1和X2的样本数量一致
X1, X2 = X1.align(X2, join='inner')

# 创建一个Pipeline
pipeline = Pipeline(steps=[('preprocessor', preprocessor)])

# 现在可以安全地使用pipeline进行模型训练