Automatic Summarization Based on Principal Component Analysis

Abstract

This paper describes an automatic summarization approach that constructs a summary by extracting sentences that are likely to represent the main theme of a document. The approach takes advantage of the co-occurrence relationships between words in the document to detect significant sentences. The particular technique used is Principal Component Analysis (PCA) which is one of the multivariate statistical methods. The PCA can quantify both word frequency and co-occurrence in the document on the basis of an eigenvector and its corresponding eigenvalue. We extract thematic words by performing the quantification method, and select significant sentences using the thematic words.

Experimental results using newspaper articles show that the proposed method is superior to the methods that use word frequency or lexical chain using a thesaurus.