Unknown But Interesting Recommendation Using Social Penetration

• Main Authors: Chih-Chin Chung, 鍾沚芹
• Other Authors: Ching-Lueh Chang
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/kgkuen

碩士 === 元智大學 === 資訊工程學系 === 103 === With the recent rise in popularity of social networks, millions of users have accommodated social network websites into their daily lives. Based on the social information from the social network websites, recommendation systems can predict users’ preferences and recommend the items to users. Traditional recommendation systems recommend items with high popularity, familiarity, and similarity to users. Such recommendation process might encounter two problems: (1) If the recommending item is very popular, the target user may already know it; (2) The target user may not be interested in the items recommended by his/her familiar users. Recently, to improve the traditional recommendation systems, a UBI system is proposed to discover unknown but interesting items. Although the UBI system provides satisfactory recommendation performance, some important information are not completely considered. The present study expends on the UBI algorithm with considering two additional factors: (1) content-based knowledge: considering user’s interests and categories, and (2) aging effect: considering the posted time of each item. Moreover, we derived a social penetration (SP) algorithm to combine the scores of content-based knowledge and aging effect to obtain social penetration score (SPS), which is used to generate a recommendation list to the target user. Experimental results demonstrate that the proposed SP algorithm can provide satisfactory recommendation list with effectively discovering UBI items.