Knowledge Discovery from Online Reviews

Online reviews have become an important information source which are helping consumers to decide which products to buy and also sellers to understand the buying behavior of consumers. Besides, mining online reviews can help manufacturers proactively review customers’ opinions and unlock insights about the new functionality and features that the market expects. However, the large number of reviews poses a considerable challenge in the process of useful information extraction and knowledge discovery. Moreover, various forms of online reviews including numeric ratings, textual comments, pictures, and videos make it difficult for consumers to summarize all the heterogeneous information for reference. This chapter will focus on three mainstream research methods of online reviews information mining: information extraction, sentiment analysis, and text categorization. It will also introduce some advanced technologies to deal with commercial issues, such as the impact of reviews on sales or product ranking, the usefulness of reviews, etc. It finally points out the expected future of techniques and commercial applications.