A query-specific opinion summarization system

This paper introduces a query-specific opinion summarization system: QOS. When input an opinion question, the system returns a summary with relevance to the opinion and target described by the question. The system has several modules to be able to do this: a question analysis and query reformulation module, a latent semantic indexing based sentence scoring module, a sentence polarity detection module, and a redundancy removal module. The system was tested on the standard dataset provided by NIST. The results ranked in the median and our system show some competence in this new field.

• Type: Conference
• Year: 2009
• Where: IEEEICCI
• Authors: Feng Jin, Minlie Huang, Xiaoyan Zhu
• Keywords: IEEEICCI 2009, Informatics, Opinion Summarization System, Query Reformulation Module, Sentence Scoring Module