Python 交叉验证后分层交叉验证和分层折叠。分层折叠弃用

我遵循了3年前的一些示例脚本，遇到了一个使用不推荐的函数的函数定义（cross_validation.StratifiedKFold）

以下是3年前的原始代码片段：

def stratified_cv(X, y, clf_class, shuffle=True, n_folds=10, **kwargs):
    stratified_k_fold = cross_validation.StratifiedKFold(y, n_folds=n_folds, shuffle=shuffle)
    y_pred = y.copy()
    # ii -> train
    # jj -> test indices
    for ii, jj in stratified_k_fold: 
        X_train, X_test = X[ii], X[jj]
        y_train = y[ii]
        clf = clf_class(**kwargs)
        clf.fit(X_train,y_train)
        y_pred[jj] = clf.predict(X_test)
    return y_pred

我尝试通过以下关于sklearn.model_selection.StratifiedKFold（）的文档对其进行更新，到目前为止，我已经做到了以下几点：

## Attempt to modernize with StratifiedKFold from sklearn.model_selection
def stratified_cv(X, y, clf_class, shuffle=True, n_splits=10, **kwargs):
    stratified_k_fold = StratifiedKFold(n_splits=n_splits)
    y_pred = y.copy()
    # ii -> train
    # jj -> test indices
    for ii, jj in stratified_k_fold: 
        X_train, X_test = X[ii], X[jj]
        y_train = y[ii]
        clf = clf_class(**kwargs)
        clf.fit(X_train,y_train)
        y_pred[jj] = clf.predict(X_test)
    return y_pred

然后，我尝试运行以下块并遇到后续错误：

print('Gradient Boosting Classifier:  {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, ensemble.GradientBoostingClassifier))))
print('Support vector machine(SVM):   {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, svm.SVC))))
print('Random Forest Classifier:      {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, ensemble.RandomForestClassifier))))
print('K Nearest Neighbor Classifier: {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, neighbors.KNeighborsClassifier))))
print('Logistic Regression:           {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, linear_model.LogisticRegression))))

错误：

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-122-a61be22f8ca9> in <module>
----> 1 print('Gradient Boosting Classifier:  {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, ensemble.GradientBoostingClassifier))))
      2 print('Support vector machine(SVM):   {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, svm.SVC))))
      3 print('Random Forest Classifier:      {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, ensemble.RandomForestClassifier))))
      4 print('K Nearest Neighbor Classifier: {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, neighbors.KNeighborsClassifier))))
      5 print('Logistic Regression:           {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, linear_model.LogisticRegression))))

<ipython-input-121-e373d74b2cca> in stratified_cv(X, y, clf_class, shuffle, n_splits, **kwargs)
      5     # ii -> train
      6     # jj -> test indices
----> 7     for ii, jj in stratified_k_fold:
      8         X_train, X_test = X[ii], X[jj]
      9         y_train = y[ii]

TypeError: 'StratifiedKFold' object is not iterable

---------------------------------------------------------------------------
TypeError回溯（最近一次调用上次）
在里面
---->1打印（'Gradient Boosting Classifier:{.2f}'。格式（度量.精度\分数（y，分层\ cv（X，y，集合.GradientBoostingClassifier）））
2打印（'Support vector machine（SVM）：{.2f}'。格式（度量.准确度\分数（y，分层\ cv（X，y，SVM.SVC）））
3打印（'Random Forest Classifier:{.2f}'。格式（度量.精度\分数（y，分层\ cv（X，y，ensemble.RandomForestClassifier）））
4打印（'K最近邻分类器：{.2f}'。格式（度量.准确度\分数（y，分层\ cv（X，y，邻居.KNeighborsClassifier）））
5打印（'Logistic回归：{.2f}'。格式（指标、准确度、得分（y、分层、线性、模型、Logistic回归）））
在分层循环中（X、y、clf类、洗牌、n分裂、**kwargs）
5#ii->列车
6#jj->测试指标
---->7对于ii，层状褶皱中的jj：
8列X_列车，X_试验=X[ii]，X[jj]
9 y_列车=y[ii]
TypeError:“StratifiedKFold”对象不可编辑

您需要使用StratifiedKFold来分割数据，而不需要对代码进行太多更改，下面应该可以使用：

from sklearn.model_selection import StratifiedKFold
from sklearn import datasets
from sklearn import metrics
from sklearn import svm

iris = datasets.load_iris()
X = iris.data
y = iris.target

def stratified_cv(X, y, clf_class, shuffle=True, n_splits=10, **kwargs):
    stratified_k_fold = StratifiedKFold(n_splits=n_splits)
    y_pred = y.copy()

    for ii,jj in stratified_k_fold.split(X, y):
            
        y_train = y[ii]
        X_train, X_test = X[ii], X[jj]
        clf = clf_class(**kwargs)
        clf.fit(X_train,y_train)
        y_pred[jj] = clf.predict(X_test)
    return y_pred
    
print('Gradient Boosting Classifier:  {:.2f}'.format(metrics.accuracy_score(y, stratified_cv(X, y, svm.SVC))))

---

这很有魅力！非常感谢：D