I can't understand LSTM's prediction output

Question

I'm studying neural networks, and I took some sample codes to debug ... I would like to understand, how can I convert a float type output to binary, because when I try to make the forecast, I know the chance of the event occurring, but I don't understand which of the events will happen, in the case 0 or 1 .

This is my entry for LSTM, a 3-dimensional matrix:

[[-0.11366585  0.06156679  0.00605332  ...  -1.02887191   1.458945    -1.00647382]
 [ 0.89127279  0.39238284  0.03161757  ...   0.97193829   1.52031675  -0.76247354]]
...
[[ 3.09754916  0.15064888   2.71586375  ...  0.97193829  -0.03852574   0.54641066]
 [-0.08069378  0.6887738   -0.29432661  ...  0.97193829   0.02207886  -0.37703]]

This is the data output:

[[0.36732605 0.6326739 ]
 [0.3584979  0.64150214]
 [0.5920879  0.40791208]
 ...
 [0.5283355  0.47166446]
 [0.5267493  0.4732507 ]
 [0.5926927  0.4073073 ]]

how it should be:

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

In this case, the correct answers are those in the right column (>) and both 0 and 1 can be between 0% to 100%, it's not because it's below 50% that the answer would be 0

And this is the code of my model:

model = Sequential()
model.add(LSTM(32, input_shape=(X.shape[1], X.shape[2]), return_sequences=True))
model.add(Dropout(0.2))

# second layer
model.add(LSTM(32, return_sequences=False))
model.add(Dropout(0.2))

# fourth layer and output
model.add(Dense(16, activation='relu'))
model.add(Dense(2, activation='softmax'))

# compile layers
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])

When I use the prediction function it only gives me the percentages, but it does not indicate whether it will be 0 or 1, how can I solve this?

Answer 1

You are getting the probabilities of each class 0 and 1. Do the following:

import numpy as np
np.argmax(output,axis=1)

If you are using Keras you can also:

model.predict_classes(X_train)