| Summary: | 碩士 === 國立高雄大學 === 資訊管理學系碩士班 === 100 === In recent years, with the growth of Internet technology and virtual community, the uses of virtual community not only play as the information receiver but also very important role to provide information. However, information overload has becoming a very serious problem and how to find information efficiently is also an important issue. In this research, we believe that users in a virtual community may affect each other, especially those with high influence.
Therefore, we observe the biggest virtual community of movies named IMDb(The Internet Movie Database),and we propose an architecture combine with the Social Network analysis and the features(Useful-index) of the IMDb to discover those users who has high influence in a virtual community.
We collected the data of 19 months (2010/1 to 2011/5) , about 17,366 users and 243,074 reviews. By applying the method we proposed, finally about 22 key users and 111 reviews were discovered. We use the box office number to certify our results, and simulated many of different situations to observe the react from our system.
|
|---|
