一、划分方式速查表（核心选型）
| KFold | 回归任务、类别均衡分类 | 均分K份，无分层，可打乱 | ⭐⭐⭐ |
| StratifiedKFold | 所有分类任务（含类别不平衡） | 分层保类别比例，标准K折交叉验证 | ⭐⭐⭐⭐⭐ |
| ShuffleSplit | 通用数据集，自定义训练/测试占比 | 随机抽样划分，无分层 | ⭐⭐⭐ |
| StratifiedShuffleSplit | 分类任务 + 自定义测试集比例 | 分层+随机打乱，单划分等价分层切集 | ⭐⭐⭐⭐⭐ |
| GroupKFold / LeaveOneGroupOut | 带分组数据（用户/样本/设备分组） | 同组数据不跨训练/测试集，防数据泄露 | ⭐⭐⭐⭐ |
| TimeSeriesSplit | 时序数据（股价、流量、时序预测） | 禁止打乱，严格按时间顺序划分 | ⭐⭐⭐⭐ |
二、场景快速判断口诀
三、全套最简可运行代码模板
统一导入
import numpy as np
from sklearn.model_selection import (
KFold, StratifiedKFold,
ShuffleSplit, StratifiedShuffleSplit,
GroupKFold, TimeSeriesSplit
)
from sklearn.utils.validation import check_random_state
# 统一随机种子（sklearn 标准用法）
seed = 42
rng = check_random_state(seed)
1. KFold（回归/均衡分类）
kf = KFold(n_splits=5, shuffle=True, random_state=rng)
for train_idx, test_idx in kf.split(X):
X_train, X_test = X[train_idx], X[test_idx]
2. StratifiedKFold（分类K折，最常用）
skf = StratifiedKFold(n_splits=5, shuffle=True, random_state=rng)
for train_idx, test_idx in skf.split(X, y):
X_train, X_test = X[train_idx], X[test_idx]
y_train, y_test = y[train_idx], y[test_idx]
3. ShuffleSplit（通用随机划分，无分层）
ss = ShuffleSplit(n_splits=5, test_size=0.2, random_state=rng)
for train_idx, test_idx in ss.split(X):
X_train, X_test = X[train_idx], X[test_idx]
4. StratifiedShuffleSplit（分类分层随机划分）
# n_splits=1 单次划分，等价分层train_test_split
sss = StratifiedShuffleSplit(n_splits=1, test_size=0.2, random_state=rng)
for train_idx, test_idx in sss.split(X, y):
X_train, X_test = X[train_idx], X[test_idx]
y_train, y_test = y[train_idx], y[test_idx]
5. GroupKFold（分组数据）
# groups 为每组对应的标签数组
gkf = GroupKFold(n_splits=5)
for train_idx, test_idx in gkf.split(X, y, groups=groups):
X_train, X_test = X[train_idx], X[test_idx]
6. TimeSeriesSplit（时序数据，不可shuffle）
tscv = TimeSeriesSplit(n_splits=5)
for train_idx, test_idx in tscv.split(X):
X_train, X_test = X[train_idx], X[test_idx]
