Topic Modeling

In this chapter, we will cover topic modeling, or the classification of topics present in a corpus of text. Topic modeling is a very useful technique that can give us an idea about which topics appear in a document set. For example, topic modeling is used for trend discovery on social media. Also, in many cases, it is useful to do topic modeling as part of the preliminary data analysis of a dataset to understand which topics appear in it.

There are many different algorithms available to do this. All of them try to find similarities between different texts and put them into several clusters. These different clusters indicate different topics.

You will learn how to create and use topic models via various techniques with the BBC news dataset in this chapter. This dataset has news that falls within the following topics: politics, sport, business, tech, and entertainment. Thus, we know that in each case, we need to have five topic clusters. This is not going to be the...