In this chapter, we have explored different techniques for text classification using LSTM, GRU, and CNN-based networks. We also touched upon topic modeling, which is a related problem in text classification. A simple example using gensim for topic modeling was introduced. We saw some solutions to sentiment classification, review rating prediction, and spam detection using real-world datasets. This gave us an overview of how to approach text classification problems with deep learning techniques using recurrent neural networks and convolutional networks. An example using transfer learning using pre-trained word embeddings was also covered. Finally, we discussed the state-of-art techniques such as extreme multi-label classification and attention networks that can be applied in complex text classification scenarios.

In the next chapter, we will look at deep learning approaches regarding search and de-duplication for identifying similar documents.