1、回归
from sklearn.model_selection import train_test_split
train_X, test_X, train_y, test_y = train_test_split(X, y, test_size=0.25)
2、分类
X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.25, stratify=y)
说明:
1. shuffle默认True,设置为False时取出的顺序与原数据顺序相同,且stratify必须为None;
2. stratify按照原数据y的比例划分,用在分类当中,可为数组表示多层。