Making Predictions with Sequences Using Recurrent Neural Networks

In the previous chapter, we focused on Convolutional Neural Networks (CNNs) and used them to deal with image-related tasks. In this chapter, we will explore Recurrent Neural Networks (RNNs), which are suitable for sequential data and time-dependent data, such as daily temperature, DNA sequences, and customers’ shopping transactions over time. You will learn how the recurrent architecture works and see variants of the model. We will then work on their applications, including sentiment analysis, time series prediction, and text generation.

We will cover the following topics in this chapter:

• Tracking sequential learning
• Learning the RNN architecture by example
• Training an RNN model
• Overcoming long-term dependencies with Long Short-Term Memory (LSTM)
• Analyzing movie review sentiment with RNNs
• Revisiting stock price forecasting with LSTM
• Writing your own War and Peace...