Java: Data Science Made Easy

Text clustering

In Chapter 5, Unsupervised Learning - Clustering and Dimensionality Reduction, we covered dimensionality reduction and clustering. We already discussed how to use dimensionality reduction for texts, but have not yet spoken about clustering.

Text clustering is also a useful technique for understanding what is a collection of documents. When we want to cluster texts, the goal is similar to non-text cases: we want to find groups of documents such that they have a lot in common: for example, the documents within such group should be on the same topic. In some cases, this can be useful for IR systems. For example, if a topic is ambiguous, we may want to group the search engine results.

K-means is a simple, yet powerful clustering ...