A Graphic Reputation Analysis System for Mining Japanese Weblog Based on both Unstructured and Structured Information

In today's rapidly expanding information environment, an immediate need exists for better tools to help a user dynamically navigate large result lists and search structures. However, at present a lot of research is focused only on the information analysis step, and is not continued on to the step that makes the analysis results visible, especially for Asian languages. In contrast, we propose a novel graphic reputation analysis system for Japanese. It not only integrates the analysis of unstructured opinion information and structured review rating data, but also presents the information using a simple and flexible GUI with visual mining graphs. The proposed system can help inexperienced users, who want to look over opinions on the Web before purchasing something, to make a quick decision. It also gives useful hints for other Asian languages. In this paper, we describe the detailed algorithms of the system, and show the performance through experiments.