machine-learning - 如何使用 to_categorical 将 [[4,7,10],[10,20,30]] 转换为一种热编码

Question

我正在研究 LSTM。

输出是分类的。

其格式为 [[t11,t12,t13],[t21,t22,t23]

我能够为 1d 阵列做到这一点，但我发现很难为 2d 阵列做到这一点。

from keras.utils import to_categorical
print(to_categorical([[9,10,11],[10,11,12]]))

输出

[[ 0.  0.  0.  0.  0.  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.  0.  0.  0.  1.  0.]
[ 0.  0.  0.  0.  0.  0.  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.  0.  0.  0.  1.]]

有两个不同的输入，每个都有 3 个时间步长，但在输出中它们全部组合在一起。

我需要它，

[[[ 0.  0.  0.  0.  0.  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.  0.  0.  0.  1.  0.]],

[[ 0.  0.  0.  0.  0.  0.  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.  0.  0.  0.  1.]]]

标签： machine-learningkerasdeep-learning

Answer 1

我意识到我可以通过重塑来实现我想要的，

print(a.reshape(2,3,13))



[[[ 0.  0.  0.  0.  0.  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.  0.  0.  0.  1.  0.]]

[[ 0.  0.  0.  0.  0.  0.  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.  0.  0.  0.  1.]]]