| Summary: | 碩士 === 國立清華大學 === 科技管理研究所 === 95 === Online communities allowing users to express personal opinions and preferences (e.g., products or product features they like or dislike) are becoming increasingly popular in recent years. However, due to the openness and anonymity of opinion-sharing communities, users also face a very challenging issue; that is, whether to believe or disbelieve information asserted by other users in the community. Therefore, it is desirable to develop an effective mechanism to better facilitate users’ information search and browsing process in online communities. In response, the purpose of this study is to develop a data-mining-based reputation prediction approach for predicting reputation of reviewers in opinion-sharing communities. The prediction of reputation scores of members in an opinion-sharing community can help users to easily find reputable reviewers in the community and can facilitate users to judge whether to believe or disbelieve reviews written by different reviewers in the community. In this study, we identify fourteen independent variables related to review, rating, and trust activities/behaviors of members and apply M5 and SVM Regression as our underlying learning algorithm. Our empirical evaluation results on the basis of four product categories suggest that our proposed approach can satisfactorily predict reputation scores of members in opinion-sharing communities. In addition, M5 appears to be more effective than SVM Regression. Our empirical evaluation also shows that trust-related variables play important roles in predicting reputation scores of members in an opinion-sharing community.
|
|---|
