Review Summarization Based on Linguistic Knowledge

Review Summarization Based on Linguistic Knowledge
리뷰 마이닝; 자연어 처리; 리뷰 요약; Domain-knowledge extraction; aspect-opinion extraction; linguistic knowledge; review summarization
Issue Date
International Workshop on Social Networks and Social Web Mining
, 105-114
In this paper, we present a review summarization method based on linguistic knowledge in order to generate a relevant summary from the product or service review. The review summarization task can be divided into two subtasks: the domain-knowledge extraction which identifies clue words through statistical techniques, and the aspect-opinion extraction which determines positive/negative aspect-opinion pairs for a new sentence. Through checking the binary association in a given re- view, our system can automatically acquire an aspect-opinion list and the ranked sentences including aspect-opinion pairs. In this study, in order to appropriately represent a review summary, we try to incorpo- rate linguistic knowledge into two steps of review summary reporting. In the first step, we reduce the number of the aspect-opinion pairs by using semantic similarity between clue words. In the second step, we esti- mate the sentence scores by utilizing several features such as the number of aspects and the number of opinions. Experimental results show that the performance of the review summarization task can be improved by utilizing such linguistic knowledge. As an experimental data, we utilize Korean reviews of a restaurant domain and a movie domain.
Appears in Collections:
KIST Publication > Conference Paper
Files in This Item:
There are no files associated with this item.
RIS (EndNote)
XLS (Excel)


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.