基于向量空间模型的在线拍卖商品推荐  

Recommendation of online auctioning items based on vector space model

作　　者：李雪峰[1] 刘鲁[1] 张曌[1] 

机构地区：[1]北京航空航天大学经济管理学院

出　　处：《辽宁工程技术大学学报（自然科学版）》2006年第2期272-275,共4页Journal of Liaoning Technical University (Natural Science)

基　　金：国家自然科学基金资助项目(70371004);高等学校博士学科点专向科研基金资助项目(20040006023)

摘　　要：随着在线拍卖的迅速发展,网站的用户数和商品数急剧增加,使得信息过载和如何提高客户忠诚度的问题凸现出来。为了解决这一问题,采用理论分析和实验的方法,对用户在拍卖网站的行为进行了详细分析,提出了偏好序的概念,在此基础上建立了基于向量空间的用户偏好模型和拍卖商品模型,然后利用内容过滤技术进行拍卖商品的个性化推荐。研究结果表明,买者偏好模型的权重矩阵形式很好地克服了传统的向量表示形式,全面地表示了买者的偏好。研究使得目前拍卖网站的推荐手段更加丰富,提高了推荐的个性化、智能化程度。With the growth of online auction, the ever-increasing users＇ size and auctioned goods cause the problem of information overload, how to enhance the customer loyalty becomes a critical issue facing by most online auction websites. To overcome the problem, theory analysis and experiment methods are used. This paper analyzes the user behaviors on the auctioned website in detail and addresses the concept of preference order. Then, the user preference model and auction goods model are constructed based on vector space. Thus, the content filtering technology is used to make the personalized recommendation of goods. The research result shows that weight matrix is prior to vector form, and make recommendation instruments of auction websites more plentiful. Further, the result promotes personalization and intelligence degree.

关 键 词：顾客对顾客 向量空间模型 用户偏好模型 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]