Extracting Experiences Using Dependency Parsing on Japanese e-Commerce Websites

Abstract

In recent years, the expansion of e-commerce has led to a rapid increase in the number of customer reviews on websites. In general, the reviews are written by consumers after they have purchased or used the products. However, reviews may not only be written by users who have experience of the products, making it difficult to determine which reviews are useful references when purchasing items. On e-commerce websites, in particular, reviews written by consumers who have actually used or purchased the items are useful to both consumers and shops, so a method of extracting such experiential information is needed. In this paper, we propose a method for extracting reviews that contain useful information using Japanese dependency parsing. Our method extracts rather less information than previous methods, but requires less processing to achieve almost the same accuracy as conventional approaches.