如何将numpy字符串数组转换为numpy数组，以使用MultiLabelBinarizer？

I have a numpy array like that:

我有一个像这样的numpy数组:

l1 = (['United States', 'England', 'South Africa']).

Sometimes it could have >1 Value:

有时它可能有> 1值:

l1 = ([['United States','South Korea'], 'England', 'South Africa'])

I want to use MultiLabelBinarizer to encode these values. According to the documentation of fit_transform in the scikit-learn documentation. The parameter should be

我想使用MultiLabelBinarizer来编码这些值。根据scikit-learn文档中的fit_transform文档。参数应该是

y : iterable of iterables A set of labels (any orderable and hashable object) for each sample. If the classes parameter is set, y will not be iterated.

y:iterable of iterables每个样本的一组标签(任何可订购和可散列对象)。如果设置了classes参数,则不会迭代y。

How can I convert this numpy array of list and single strings into sets?

如何将这个numpy列表和单个字符串数组转换为集合?

I have tried this:

我试过这个:

value = [set(v) for v in l1]
list_2sets = np.asarray(value)

But it seems it doesn't work properly.

但它似乎无法正常工作。

The thing is that I do not have the values considered (all countries). If I had this, I tried the following which works:

问题是我没有考虑价值(所有国家)。如果我有这个,我尝试了以下工作:

 mlb.fit_transform(headings.split(', ') for headings in l1)

Being headings the list of all values considered:

作为标题列出所有考虑的价值观:

['England','Spain', ...]

But I do not have those values so far, so I wanted to try applying MLB without 'headings'

但到目前为止我还没有这些价值观,所以我想尝试在没有“标题”的情况下应用MLB

1 个解决方案

#1

Try to preprocess your array of strings as follows:

尝试预处理字符串数组,如下所示:

In [50]: l1 = [[x] if isinstance(x, (str)) else x for x in l1]

In [51]: l1
Out[51]: [['United States', 'South Korea'], ['England'], ['South Africa']]

For Python 2.x:

对于Python 2.x:

In [50]: l1 = [[x] if isinstance(x, (str, unicode)) else x for x in l1]

In [51]: l1
Out[51]: [['United States', 'South Korea'], ['England'], ['South Africa']]

#1