Chinese Multi-document Summarization Based on Opinion Similarity

Abstract

This paper proposes a strategy for Chinese review multi-document summarization based on opinion extraction and opinion similarity. Firstly, sentences are sorted according to their weights which are calculated based on word frequency statistics. Then, it extracts opinions from sentences through finding out sentiment features and corresponding topic features. After that, opinion similarity calculation method is introduced. Based on opinion similarity calculation, sentences which contain more important and less redundant information are optimized as summary.