Adapted Recommendation System Based On User Browsing Behavior

• Main Authors: He-Tsun Chi, 紀和村
• Other Authors: Tung-Shou Chen
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/39c3m2

碩士 === 臺中技術學院 === 資訊工程系碩士班 === 99 === To date, the internet development has matured where information and knowledge has also entered the digital age. The traditional learning behavior has expanded from entity study to virtual learning environment which makes the transmission of information and knowledge diversified. Users are also using surfing to absorb various types of knowledge on the internet. However, the information technologies are changing fast with time. Many big websites provide rich database of knowledge which resulted in information explosion. Therefore, many experts and scholars are using various data mining methods to solve the problems of information explosion by filtering information. This paper proposed a recommendation system based on web log of the website users where three methods were used to recommend information. First, we used association rule to analyze users’ behaviors on the web to generate meaningful rules and to proceed with the recommendation of behavior pattern. The best path is generated after transferring association rules from click-select. Finally, the clustering rules on users in the clusters to recommend. This paper also search for the support degree of the different users between rules, and proposed a new concept at the time interval of the users’ browsing session; the time interval of the users browsing session is different from various types of websites. The users’ browsing behaviors, paths clicked most and users recommended clustering results were used to adapt a recommendation list for users. The recommended list can increase the users interest on the items offered on the web pages.