A Collaborative Recommendation Mechanism based on Time-Variant Participation

• Main Authors: Yung-Fang Hung, 洪永芳
• Other Authors: Hsing Mei
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/71160910421679749621

碩士 === 輔仁大學 === 資訊工程學系 === 98 === Due to the rapid growth of web information, it is difficult to provide appropriate recommendation to users in such an information overload environment. A recommendation system should provide personalized recommendations to help users to obtain the necessary information and services effectively. However, the traditional recommendation systems typically use the user's explicit preferences as the main information source. A large amount of objects and a small amount of rating information are likely to cause the sparsity problem. Also, a new user without any available rating information could encounter the cold start problem. Most of all, a timely recommendation is not possible without considering the time factor. This thesis proposes a recommendation mechanism that considers the user’s participation and time factor. The proposed recommendation mechanism is based on Collaborative Filtering and considers the user’s Time-Variant Participation within a community. The participation and contribution of user in the community is calculated to help the sparsity problem. By recommending the highest participated member in a community could solve the cold start problem. Through the degrading consideration of participation and contribution by time, more appropriate recommendations are given. This recommendation mechanism has been validated on the Social Learning Space (SLS).