Text Summarization Using Semantic Technique

This work proposes semantic technique as a new approach for text summarization of
online news/ journal articles. This text summarization project contains two part: parsing
and semantic analysis. At first, for the parsing, we set up criteria to evaluate importance
of sentences within text such as position, length. According to this, only sentences with
high score of importance will be selected. The number of sentence selected depends on
how much compact that users expect summary output would be. System combines those
sentences into summary draft. Second part is semantic analysis; the technique we use
here is lexical semantic approach. During this part, system will use Word Net lexical
database to analyze words within summary draft. This database had words linked
through semantic relations such as synonym, antonym, hyponymy, and more. Our
project would end up by evaluation process. We evaluate accuracy of summary output,
also compare and find differences between human summary work and system result.
4