Python 使用 Keras np_utils.to_categorical 的问题
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/41494625/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Issues using Keras np_utils.to_categorical
提问by Eduardo
I'm trying to make an array of one-hot vector of integers into an array of one-hot vector that keras will be able to use to fit my model. Here's the relevant part of the code:
我正在尝试将整数的 one-hot 向量数组转换为 keras 能够用于拟合我的模型的 one-hot 向量数组。这是代码的相关部分：
Y_train = np.hstack(np.asarray(dataframe.output_vector)).reshape(len(dataframe),len(output_cols))
dummy_y = np_utils.to_categorical(Y_train)
Below is an image showing what Y_trainand dummy_yactually are.
下面是一张图片，显示了什么Y_train和dummy_y实际上是什么。
I couldn't find any documentation for to_categoricalthat could help me.
我找不到任何to_categorical可以帮助我的文档。
Thanks in advance.
提前致谢。
回答by yardstick17
np.utils.to_categoricalis used to convert array of labeled data(from 0 to nb_classes-1) to one-hot vector.
np.utils.to_categorical用于将标记数据数组（从 0 到 nb_classes-1）转换为单热向量。
The official doc with an example.
带有示例的官方文档。
In [1]: from keras.utils import np_utils # from keras import utils as np_utils
Using Theano backend.
In [2]: np_utils.to_categorical?
Signature: np_utils.to_categorical(y, num_classes=None)
Docstring:
Convert class vector (integers from 0 to nb_classes) to binary class matrix, for use with categorical_crossentropy.
# Arguments
y: class vector to be converted into a matrix
nb_classes: total number of classes
# Returns
A binary matrix representation of the input.
File: /usr/local/lib/python3.5/dist-packages/keras/utils/np_utils.py
Type: function
In [3]: y_train = [1, 0, 3, 4, 5, 0, 2, 1]
In [4]: """ Assuming the labeled dataset has total six classes (0 to 5), y_train is the true label array """
In [5]: np_utils.to_categorical(y_train, num_classes=6)
Out[5]:
array([[ 0., 1., 0., 0., 0., 0.],
[ 1., 0., 0., 0., 0., 0.],
[ 0., 0., 0., 1., 0., 0.],
[ 0., 0., 0., 0., 1., 0.],
[ 0., 0., 0., 0., 0., 1.],
[ 1., 0., 0., 0., 0., 0.],
[ 0., 0., 1., 0., 0., 0.],
[ 0., 1., 0., 0., 0., 0.]])
回答by Pranzell
from keras.utils.np_utils import to_categorical
to_categorical(0, max_value_of_array)
It assumes the class values were in string and you will be label encoding them, hence starting everytime from 0 to n-classes.
它假设类值在字符串中，并且您将对它们进行标签编码，因此每次从 0 到 n 类开始。
for the same example:- consider an array of {1,2,3,4,2}
The output will be [zero value, one value, two value, three value, four value]
array([[ 0., 1., 0., 0., 0.],
[ 0., 0., 1., 0., 0.],
[ 0., 0., 0., 1., 0.],
[ 0., 0., 0., 0., 1.],
[ 0., 0., 1., 0., 0.]],
Let's look at another example:-
让我们再看一个例子：-
Again, for an array having 3 classes, Y = {4, 8, 9, 4, 9}
to_categorical(Y) will output
array([[0., 0., 0., 0., 1., 0., 0., 0., 0., 0. ],
[0., 0., 0., 0., 0., 0., 0., 0., 1., 0. ],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 1. ],
[0., 0., 0., 0., 1., 0., 0., 0., 0., 0. ],
[0., 0., 0., 0., 0., 0., 0., 0., 0., 1. ]]
