Neural Networks with Keras Cookbook

How to do it...

The steps we'll adopt are as follows (the code file is available as RNN_and_LSTM_sentiment_classification.ipynb in GitHub):

1. Define the model. The only change from the code we saw in Implementing RNN for sentiment classification recipe will be the change from simpleRNN to LSTM in the model architecture part (we will be reusing the code from step 1 to step 6 in the Implementing RNN for sentiment classification recipe):

embedding_vecor_length=32max_review_length=26model = Sequential()model.add(Embedding(input_dim=12533, output_dim=32, input_length = 26))

The input for the embedding layer is the total number of unique IDs present in the dataset and the expected dimension to which each word needs to be converted (output_dim