Multiple propositions for multiple class prediction - keras

Question

I am working on a word prediction problem. I have examples of career path, and I would like to be able to predict a next person's job using their last 2 jobs. I have built a LSTM model to perform it
I have a problem when intenting to get multiple results from keras model.predict_classes function. It only returns 1 result. I would like to get multiple results, ordered by their probability.
Here is the code :
from numpy import array
from keras.preprocessing.text import Tokenizer
from keras.utils import to_categorical
from keras.preprocessing.sequence import pad_sequences
from keras.models import Sequential
from keras.layers import Dense
from keras.layers import Dropout
from keras.layers import LSTM
from keras.layers import Embedding
# generate a sequence from a language model
def generate_seq(model, tokenizer, max_length, seed_text, n_words):
in_text = seed_text
# generate a fixed number of words
for _ in range(n_words):
# encode the text as integer
encoded = tokenizer.texts_to_sequences([in_text])[0]
# pre-pad sequences to a fixed length
encoded = pad_sequences([encoded], maxlen=max_length, padding='pre')
# predict probabilities for each word
yhat = model.predict_classes(encoded, verbose=1)
print('yhat = ' + yhat)
#print('yhat : ' + str(yhat))
# map predicted word index to word
out_word = ''
for word, index in tokenizer.word_index.items():
if index == yhat:
out_word = word
break
# append to input
in_text += ' ' + out_word
return in_text
# source text
data = """apprenti electricien chefOdeOprojet \n
soudeur chefOdeOsection directeurOusine\n
mecanicien chefOdeOsection directeurOadjoint\n
ingenieur chefOdeOprojet directeurOadjoint directeurOusine\n
ingenieur chefOdeOprojet \n
apprenti soudeur chefOdeOsection chefOdeOprojet\n
ingenieurOetude chefOdeOprojet\n
ingenieurOetude manager chefOdeOprojet directeurOdepartement\n
apprenti gestionOproduction manager directeurOdepartement\n
ingenieurOetude commercial\n
soudeur ingenieurOetude manager directeurOadjoint\n
ingenieurOetude directeurOdepartement directeurOusine\n
apprenti soudeur\n
agentOsecurite chefOsecurite\n
apprenti mecanicien ouvrier manager\n
commercial directeurOadjoint\n
agentOsecurite chefOsecurite\n
directeurOusine retraite\n
ouvrier manager\n
ingenieur vente\n
secretaire comptable\n
comptable chefOcomptable\n
chefOcomptable directeurOdepartement\n
assistant secretaire comptable\n
assistant comptable\n
assistant secretaire commercial\n
commercial chefOdeOprojet\n
commercial vente chefOdeOprojet\n
electricien chefOdeOsection\n
apprenti ouvrier chefOdeOsection\n"""
# integer encode sequences of words
tokenizer = Tokenizer()
tokenizer.fit_on_texts([data])
encoded = tokenizer.texts_to_sequences([data])[0]
# retrieve vocabulary size
vocab_size = len(tokenizer.word_index) + 1
print('Vocabulary Size: %d' % vocab_size)
# encode 2 words -> 1 word
sequences = list()
for line in data.split('\n'):
encoded = tokenizer.texts_to_sequences([line])[0]
for i in range(2, len(encoded)):
sequence = encoded[i-2:i+1]
sequences.append(sequence)
print('Total Sequences: %d' % len(sequences))
# pad sequences
max_length = max([len(seq) for seq in sequences])
sequences = pad_sequences(sequences, maxlen=max_length, padding='pre')
print('Max Sequence Length: %d' % max_length)
# split into input and output elements
sequences = array(sequences)
X, y = sequences[:,:-1],sequences[:,-1]
y = to_categorical(y, num_classes=vocab_size)
# define model
model = Sequential()
model.add(Embedding(vocab_size, 10, input_length=max_length-1))
model.add(LSTM(50))
model.add(Dropout(0.2))
#model.add(Dense(units = 3, activation = 'relu'))
model.add(Dense(vocab_size, activation='softmax'))
print(model.summary())
# compile network
model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
# fit network
model.fit(X, y, epochs=500, verbose=0)
# evaluate model
print(generate_seq(model, tokenizer, max_length-1, 'electricien secretaire', 1))
and there is the console display:
Vocabulary Size: 24
Total Sequences: 20
Max Sequence Length: 3
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
embedding_2 (Embedding) (None, 2, 10) 240
_________________________________________________________________
lstm_2 (LSTM) (None, 50) 12200
_________________________________________________________________
dropout_2 (Dropout) (None, 50) 0
_________________________________________________________________
dense_2 (Dense) (None, 24) 1224
=================================================================
Total params: 13,664
Trainable params: 13,664
Non-trainable params: 0
_________________________________________________________________
None
1/1 [==============================] - 0s 86ms/step
yhat = [1]
electricien secretaire chefodeoprojet

Answer 1

If I understand your question correctly, you would like to see the probabilities associated with each class of a multi-classification problem?
The code looks pretty correct to me, but I would recommend trying a different evaluation step. I have gotten multi-classifier outputs with the following snippet:
# Fit the model
print "Fitting model..."
model.fit(np.asarray(self.X), np.asarray(self.Y), epochs=200, batch_size=10)
print "Model fitting complete."
self.TEST = np.asarray(self.TEST).reshape(( test_data.shape[0], 1, 128))
print "Predicting on Test (unseen) data..."
predictions = model.predict(self.TEST)
# Sigmoid predictions
labels = np.zeros(predictions.shape)
labels[predictions>0.5] = 1
print "Prediction labels for unseen: " + str(labels)
The output:
Prediction labels for unseen:
[[ 0. 1. 0. 0.]
[ 0. 1. 0. 0.]
[ 0. 1. 0. 0.]
[ 0. 1. 0. 0.]
[ 0. 1. 0. 0.]
[ 0. 0. 1. 0.]
[ 0. 0. 1. 0.]
[ 0. 0. 1. 0.]]
Each row denotes the classification of one sample; the index of the 1 represents which class (A,B,C,D) the sample fell into.