SentiSummary: Sentiment Summarization for User Product Reviews

This paper describes an approach to summarize and analyze user product reviews using summarization techniques found in multi-document summarization as well as sentiment anal-ysis to extract opinions found in text. Previously, most re-search has focused on creating summaries for formal text such as The Wall Street Journal or other major news out-let. However, those same summarizers do not perform well on informal language such as blogs and user reviews. Our system uses sentiment analysis to analyze user product re-views for multiple products, identify and parse the positive and negative viewpoints across multiple reviews, summarize those viewpoints and display the aggregated information in a user friendly and useful manner. 1