一种基于内容过滤的服务资源推荐新技术  

A New Recommendation Technology of Service Resources Based on Content Filtering

作　　者：刘平峰[1,2] 杨柳[1,2] 朱孔真[1,2] 

机构地区：[1]武汉理工大学经济学院,湖北武汉430070 [2]武汉理工大学电子商务与智能服务研究中心,湖北武汉430070

出　　处：《武汉理工大学学报（信息与管理工程版）》2014年第5期668-672,共5页Journal of Wuhan University of Technology：Information & Management Engineering

基　　金：国家科技支撑计划基金资助项目(2013BAH13F01;2012BAH93F04);教育部人文社科基金资助项目(09YJCB70025);教育中央高校基本业务专项资金资助项目(2012-IB-060);中央高校基本科研业务费专项资金资助项目(2012-JL-03);武汉理工大学自主创新基金资助项目(2014-ZY-175)

摘　　要：针对传统基于内容过滤的服务资源推荐技术中,用户对不同属性偏好权重的计算依赖于用户Web使用行为,不能适用于用户行为数据缺乏的场合,以及用户对属性值偏好计算缺乏用户评分反馈考虑等问题,提出了一种基于内容过滤的服务资源推荐新技术,该技术依据用户消费业务数据和评分反馈,通过计算服务资源属性的信息熵给属性权重赋值,并根据属性值变化频率对属性权重进行调整,得到用户对属性的偏好权重;根据属性值概率分布和用户评分,引入遗忘因子计算用户对属性值的偏好。实验结果表明,该技术建立的用户偏好模型和依据该推荐的服务资源结果更为准确。For the traditional recommendation technology of service resources based on content filtering , the calculation of the user＇s preference weights for different attribute depends on the user＇s past behavior in the Web .However , it is not appropriate for the situation lacking the user＇s past behavior data and the user＇s rating feedback when calculating the user＇s preferential attribute value.A new recommendation technology of service resources was proposed based on content filtering .It can obtain the user＇s preferential attribute weight , through the business data of the user＇s consumption and the user＇s rating feedback , assign attribute weights based on the information entropy of the service resource attributes , and adjust the attribute weights according to the change frequency of attribute values .The preference to the attribute value was calculated based on the probability distribution to the attribute values, the user＇s ratings, and the forgetting factor.The results show that the user＇s preference model and the recom-mended service resources according to the new recommendation technology are more accurate .

关 键 词：内容过滤 服务资源推荐 消费偏好 

分 类 号：TP39[自动化与计算机技术—计算机应用技术] F270[自动化与计算机技术—计算机科学与技术]