¡@

With the steadily increasing volume of e-commerce transactions, the amount of userprovided product reviews is increasing on the Web. Because many customers feel that they can purchase product based on the experiences of others that are obtainable through product reviews, the review summarization process has become important. In particular, feature-based product review summarization is needed in order to satisfy the detailed needs of some customers. To achieve such summarization, numerous techniques using natural language processing (NLP), machine learning, and statistical approaches that can evaluate product features within a collection of review documents have been studied. Many of these techniques require sentiment analysis or feature scoring methods. However, existing sentiment analysis methods are limited when determining the sentiment polarity of context-sensitive words, and existing feature scoring methods are limited when only the overall user score is used to evaluate individual product features. In our summarization approach, context-sensitive information is used to determine sentiment polarity while opinioned-feature frequency is used to determine feature scores. Based on experiments with actual review data, our method improved the accuracy of the calculated feature scores and outperformed existing methods.

Keywords: product review summarization, opinion mining, sentiment analysis, feature scoring, user score

Retrieve PDF document (201011_03.pdf)

Received June 9, 2009; revised August 6 & October 5, 2009; accepted November 18, 2009.
Communicated by Chin-Teng Lin.
^* This research was supported by the MKE (The Ministry of Knowledge Economy), Korea, under the ITRC (Information Technology Research Center) support program supervised by the NIPA (National IT Industry Promotion Agency) grant No. NIPA-2009-C1090-0902-0031.