Concept - Document Clustering

Definition (s.34): given a set of documents and a similarity measure among documents, find clusters such that documents in one cluster are more similar to one another and documents in separate clusters are less similar, using some clustering algorithm.

Applications (s.34): topical clustering of news stories; email message thread identification; grouping of document versions.

Key question (s.34): which similarity measure is good for document vectors? Answers: Concept - Jaccard Coefficient (binary term occurrence), Concept - Cosine Similarity (TF / TF-IDF; commonly used with TF-IDF, s.38), embedding-based similarity (doc2vec, averaged Concept - Word Embeddings, word mover’s distance; s.40).

Figure s.34: two circled clusters of documents, each with a highlighted representative document.

Source: 01 Intro - Chunk 02 (s.31-48)

Related

Concept - Pattern Discovery, Concept - Document Classification, Concept - TF-IDF