Opinion Analysis for Enhanced Live Business Intelligence

An important task in Business Intelligence (BI) applications is to gather and analyze scattered opinions about an entity such as a product, a brand or a company. Thanks to Web 2.0 technology which has enabled more and more people to freely express their opinions in real time, the Web has now become an extremely valuable source for analyzing people’s opinions. However, since the opinions are usually expressed as unstructured text scattered in different sources, with the current technologies, it is still difficult for analysts to integrate and digest all opinions relevant to a specific topic. In this project, we propose to develop a novel application system for Live BI that can automatically integrate opinions from different kinds of sources and support comparative sentiment analysis and real-time tracking of the opinions. The system would allow a user to register topics or entities to be monitored and compared and “watch” how the opinions evolve over time in a similar way to what people can do with the stock index. “triggers” will be supported which would generate an alert when special event happens, such as the amount of positive opinions about a competitor’s product has significantly exceeded one’s own product. Three main technical challenges fundamental to such a novel system include opinion integration, sentiment tagging, and online comparative opinion analysis. We will address these challenges generally with probabilistic approaches that can be applied to multiple domains for facilitating multiple kinds of business decisionmaking, such as for customer relationship management, marketing research, and competitive intelligence.