| Summary: | 碩士 === 中國文化大學 === 資訊管理研究所 === 98 === A large amount of data enters major websites everyday. Users usually cannot search and determine appropriate information when they come across overloaded data. Active recommendation is an answer to this problem. Through coordinated filtering analysis, appropriate information can be recommended after filtering.
That the operating behaviors of internet users can be divided into two categories : searching and browsing. Users directly search for the information when the information is specific and clear; when unclear, they would browse and search possible materials. They would accept suggestions from the Internet when they browse; thus, recommendation system can lead users to their desired data. This can increase the merchandise exposure and the purchase intention; while users’ behaviors are recorded and provided for system analysis, in order to modify future user behavior and merchandise recommendation.
Therefore, the study selected a theme website as the experimental subject and used user’s browsing history and user’s rating data sheet to construct a mechanism for recommendation systems through collaborative filtering and Slope One algorithm. This mechanism can allow users to obtain the information which meets their needs effectively and enable website operators to further offer more information services.
|
|---|
