python错误：数组的索引太多

Question

我的输入是一个导入到 postgresqldb 的 csv 文件。后来我正在使用 keras 构建一个 cnn。我下面的代码给出了以下错误“IndexError：数组索引过多”。我对机器学习很陌生，所以我不知道如何解决这个问题。有什么建议吗？

X = dataframe1[['Feature1','Feature2','Feature3','Feature4','Feature5','Feature6','Feature7','Feature8','Feature9','Feature10','Feature11\1','Feature12','Feature13','Feature14']]
Y=result[['label']]



# evaluate model with standardized dataset
results = cross_val_score(estimator, X, Y, cv=kfold)
print("Results: %.2f%% (%.2f%%)" % (results.mean()*100, results.std()*100))

错误

    ---------------------------------------------------------------------------
    IndexError                                Traceback (most recent call last)
    <ipython-input-50-0e5d0345015f> in <module>()
          2 estimator = KerasClassifier(build_fn=create_baseline, nb_epoch=100, batch_size=5, verbose=0)
          3 kfold = StratifiedKFold(n_splits=10, shuffle=True, random_state=seed)
    ----> 4 results = cross_val_score(estimator, X, Y, cv=kfold)
          5 print("Results: %.2f%% (%.2f%%)" % (results.mean()*100, results.std()*100))

    C:\Anacondav3\lib\site-packages\sklearn\model_selection\_validation.py in cross_val_score(estimator, X, y, groups, scoring, cv, n_jobs, verbose, fit_params, pre_dispatch)
        129 
        130     cv = check_cv(cv, y, classifier=is_classifier(estimator))
    --> 131     cv_iter = list(cv.split(X, y, groups))
        132     scorer = check_scoring(estimator, scoring=scoring)
        133     # We clone the estimator to make sure that all the folds are

    C:\Anacondav3\lib\site-packages\sklearn\model_selection\_split.py in split(self, X, y, groups)
        320                                                              n_samples))
        321 
    --> 322         for train, test in super(_BaseKFold, self).split(X, y, groups):
        323             yield train, test
        324 

    C:\Anacondav3\lib\site-packages\sklearn\model_selection\_split.py in split(self, X, y, groups)
         89         X, y, groups = indexable(X, y, groups)
         90         indices = np.arange(_num_samples(X))
    ---> 91         for test_index in self._iter_test_masks(X, y, groups):
         92             train_index = indices[np.logical_not(test_index)]
         93             test_index = indices[test_index]

    C:\Anacondav3\lib\site-packages\sklearn\model_selection\_split.py in _iter_test_masks(self, X, y, groups)
        608 
        609     def _iter_test_masks(self, X, y=None, groups=None):
    --> 610         test_folds = self._make_test_folds(X, y)
        611         for i in range(self.n_splits):
        612             yield test_folds == i

    C:\Anacondav3\lib\site-packages\sklearn\model_selection\_split.py in _make_test_folds(self, X, y, groups)
        595         for test_fold_indices, per_cls_splits in enumerate(zip(*per_cls_cvs)):
        596             for cls, (_, test_split) in zip(unique_y, per_cls_splits):
    --> 597                 cls_test_folds = test_folds[y == cls]
        598                 # the test split can be too big because we used
        599                 # KFold(...).split(X[:max(c, n_splits)]) when data is not 100%

IndexError: too many indices for array

我应该以不同的方式声明数组或数据框吗？

Answer 1

注意User Guide 中的示例表明X 是二维的，而y 是一维的：

>>> X_train.shape, y_train.shape
((90, 4), (90,))

一些程序员对二维数组使用大写变量，对一维数组使用小写变量。

因此使用

Y = result['label']

而不是

Y = result[['label']]

---

我假设result 是一个熊猫数据框。当您使用诸如['label'] 之类的列列表对 Dataframe 进行索引时，将返回一个二维的子 DataFrame。如果使用单个字符串对 DataFrame 进行索引，则返回一维 Series。

---

最后，注意IndexError

IndexError: too many indices for array

在这条线上提出

cls_test_folds = test_folds[y == cls]

因为y 是二维的，所以y == cls 是一个二维布尔 数组，而test_folds 是一维的。情况类似如下：

In [72]: test_folds = np.zeros(5, dtype=np.int)
In [73]: y_eq_cls = np.array([(True, ), (False,)])
In [74]: test_folds[y_eq_cls]
IndexError: too many indices for array