Visualization Recommendation System Based on Social Network

• Main Authors: Yi-Hsiu Chang, 張意綉
• Other Authors: Chi-WeiLee
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/98239327386332484396

碩士 === 元智大學 === 資訊傳播學系 === 100 === Interpersonal relationship plays an important part in every person’s life. Many community websites try to connect users and keep them online by means of interpersonal relationship. The recommendation system is used to aid the users to obtain and filter essential information. It is wildly used in electronic commercial websites for recommending items and products. However, recommending related users is a rare feature to be found on websites. Aside from that, the recommendation results are listed sequentially with texts and images which is not user friendly. In this research, a visual recommendation system based on social network is designed and developed. With Graph API provided by Facebook, the similarities of users can be evaluated and then be used for recommending users with similar personal information. The results include: (1) highly similar users are recommended to the current user, (2) web pages that the current user might be interested are recommended, (3) user similarities and preferences are visualized, and (4) web pages that the current user might be interested are visualized. The features are developed to improve current social network websites. Finally, users are interviewed to understand if the usability and system requirements are satisfied.