Main Article Content

Abstract

This paper is the implementation of both topic model text summarization and Information retrieval model.
This paper conducted on identyfing topical coherency inside one paragraphs of an expository text. Expository
text is a kind of text which is intended for publicity such as journal, news etc. Many expository text consist of
long sequences of paragraphs with very little structural demarcation, while others consist of sequences of
paragraphs which still discuss the same topic. This paper uses vector space model to identify topical coherency
between paragraphs. Vector space model measures lexical similarity between adjacent paragraphs, with
assumption that the more similar two adjacent paragraphs are, the more likely it is that the current topic
continues. We use expository text from one week daily local newspaper “Kedaulatan Rakyat” from cultural and
educational topic. We choose 20 sample for testing our method, the result is not very good for the precision
among main idea and its paragraphs (approximately 60 percentage success and 40 failured).
Keyword: expository text, topical coherency, vector space model model

Article Details