基于改进协同过滤的个性化 Web 服务推荐方法研究<br />-《计算机技术与发展》

文章信息/Info

Title:: Research on Personalized Web Service Recommendation Based on Improved Collaborative Filtering

作者:: 徐堃1 ; 朱小柯2 ; 荆晓远3; 1. 南京邮电大学计算机学院,江苏南京 210003;
2. 武汉大学计算机学院,湖北武汉 430072;
3. 南京邮电大学自动化学院,江苏南京 210003

Author(s):: XU Kun 1 ; ZHU Xiao-ke 2 ; JING Xiao-yuan 3; 1. School of Computer,Nanjing University of Posts and Telecommunications,Nanjing 210003,China;
2. School of Computer,Wuhan University,Wuhan 430072,China;
3. School of Automation,Nanjing University of Posts and Telecommunications,Nanjing 210003,China

Keywords:: Web service recommendation; QoS prediction; users preference; collaborative filtering

摘要:: 目前基于协同过滤(collaborative filtering,CF)的 Web 服务推荐算法,使用的是 Web 服务的非功能性属性服务质量
(quality of services,QoS),但是这类方法直接使用所有用户的 QoS 数据进行预测,并没有考虑用户的个性化偏好问题,导致在相似邻居的选择阶段会得到不真实的相似度结果,进而影响 QoS 预测准确率。针对以上问题,提出了一种基于用户偏好的改进协同过滤 Web 服务推荐算法。该算法从 QoS 数据中提取出用户偏好数据,并将其作为近似邻居的选择标准,然后使用 top-k 算法确定目标用户及服务的相似邻居集合,最后联合相似邻居偏好比重,使用调和的皮尔逊相关系数算法(Pearson correlation coefficient,PCC)预测目标用户及服务的 QoS 值。实验结果表明,该算法能有效提高 QoS 预测准确率,从而提高了Web 服务推荐质量。

Abstract:: Existing Web service recommendation algorithms based on collaborative filtering uses quality of services of non-functional attribute. However,they make a prediction directly by means of QoS data from all users without considering the preferences of them,which lead to unreal similarity in selection of similar neighbors and further affect the accuracy of QoS. In view of that,we propose an improved collaborative filtering algorithm based on users preference. It takes the preferable data of users in QoS as the standard of similar neighbors,and then i-
dentifies the similar neighbor sets of target users or services by top-k algorithm. Finally,the Pearson correlation coefficient is used to predict the QoS of targets users or services in combination of preference ratio of similar neighbors. The experiment shows that the algorithm proposed can effectively improve the accuracy of QoS,thus enhancement of the recommendation quality of Web service.