sklearn.cross_validation引发IndexError：用作索引的数组必须是整数（或布尔）类型

Question

我在IndexError筹集sklearn.cross_validation时遇到了困难。我解决了这个问题，似乎是sklearn的错误。

我在ipython并行引擎上运行以下代码：

kf = cross_validation.KFold(num_data, k_fold)
for k, (train, test) in enumerate(kf):
    # do something

并发生错误

IndexError: arrays used as indices must be of integer (or boolean) type

发生的地方错误是：

def _iter_test_masks(self):
    """Generates boolean masks corresponding to test sets.

    By default, delegates to _iter_test_indices()
    """
    for test_index in self._iter_test_indices():
        test_mask = self._empty_mask()
        test_mask[test_index] = True # <============= Here error occurs
        yield test_mask

如何解决这个问题？

Answer 1

我终于解决了这个问题。

只需添加.astype('int64)

即可

def _iter_test_masks(self):
    """Generates boolean masks corresponding to test sets.

    By default, delegates to _iter_test_indices()
    """
    for test_index in self._iter_test_indices():
        test_mask = self._empty_mask()
        test_mask[test_index.astype('int64')] = True  # convert to int type
        yield test_mask