This chapter will introduce you to LSTM (Long Short Term Memory) model that is based on RNN. This model can be used to implement the sequence analysis.
You can follow the steps described below to create an LSTM model:
Step – 1: Import the necessary modules
from keras.preprocessing import sequence
from keras.models import Sequential
from keras.layers import Dense, Embedding
from keras.layers import LSTM
from keras.datasets import dset
Step – 2: Load the dataset
(model_train, model2_train), (model_test, model2_test) = dset.load_data(total_words = 3000)
Step – 3: Process the dataset
model_train = sequence.pad_sequences(model_train, maxlen=90)
model_test = sequence.pad_sequences(model_test, maxlen=90)
Step – 4: Create the model and then Compile it
Let’s create the model with the following code:
model = Sequential()
model.add(Embedding(3000, 128))
model.add(LSTM(128, dropout = 0.2, recurrent_dropout = 0.2))
model.add(Dense(1, activation = 'sigmoid'))
Now, to compile the code, follow the code given below:
model.compile(loss = 'binary_crossentropy',
optimizer = 'opt', metrics = ['accuracy'])
Step – 5: Training and Evaluation of the model
model.fit(
model_train, model2_train,
batch_size = 32,
epochs = 20,
validation_data = (model_test, model2_test)
)
Step – 6: Model Evaluation
score, acc = model.evaluate(model_test, model2_test, batch_size = 32)
print('The total test score is:', score)
print('The total test accuracy is:', acc)
You can simply follow the steps described above to test the results of the model.