基于海量数据挖掘的个性化推荐系统  

Research of personalized recommender system based on data mining on magnanimity data

作　　者：郭晔[1] 王浩鸣[1] 杨新安[1] 

机构地区：[1]西安财经学院计算机科学系,陕西西安710061

出　　处：《西北大学学报（自然科学版）》2006年第6期899-902,共4页Journal of Northwest University（Natural Science Edition）

基　　金：陕西省自然科学基金资助项目(2005F08)

摘　　要：目的建立海量数据环境中具有个性化的推荐系统。方法在普通文献推荐系统的基础上,增加基于链接页面的Pagerank计算,从而更精确地表示查询页面相对于特定用户的查询价值。结果结合了基于页面内容的查询方法与基于链接的查询方法的优点。结论具有一定的研究价值,值得在未来的研究工作中加以完善。Aim To setup a personalized recommender system based on on data mining on magnanimity data. Methods Present a personalized recommender system model combining the text categorization with the pagerank. The features of pages were extracted in order to form the feature vector, which will be used in computing the difference between the documents or keywords with the user＇s interests and the given domain. The links between the pages were divided into two parts, the inter-link and the intra-link, according to the position of pages. All links were of different weight in the link matrix. The final order of the documents was determined by the vector distance and the eigenvector of the link matrix. Results It combined the advantages of the method based on content and on links. Conclusion It is valuable to be researched in the future.

关 键 词：文档分类 特征提取 向量空间 邻接矩阵 PAGERANK 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]