Applying LDAP directory service for a personalized recommendation based on product taxonomy and user preference ontology

• Main Authors: Huan-Cheng Lin, 林恒正
• Other Authors: Li-Hua Li
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/85938218000669314066

碩士 === 朝陽科技大學 === 資訊管理系碩士班 === 95 === In this paper, we apply a personalization recommendation for E-commerce site. Personalized recommendation is used to overcome the product information overload problem. The product taxonomy is generally available for E-commerce site in managing product information. We use ontology to present user preference, and then user preference ontology is in accordance with product taxonomy structure. Web usage mining is a way to know user’s number of visits to a specific product. The hits numbers are filled in the instances of ontology. Furthermore user preference ontology are stored in LDAP directory database. It is characteristic of LDAP to have powerfully search capability. After ontology inference method, cosine similarity measure is applied when calculating the distance between two users ontology nodes vectors. Therefore we find the similar preference users for target recommend user. The recommend lists are produced by way of recommend engine. We have implemented the mechanism with JSP web page and showed the ontology tree and recommend lists for target recommend user.