论文部分内容阅读
为改善Web服务QoS预测效率,向用户提供高质量的Web服务,提出一种融合位置最近邻法则的扩展矩阵分解算法.该方法首先利用用户和服务的位置信息进行用户和服务的最近邻选择,克服了传统的QoS预测算法对Web服务位置信息利用不准确或不足的问题.然后将邻域信息融入矩阵分解框架,改善了矩阵分解技术在QoS预测中本地信息利用不足的问题,同时采用梯度下降算法进行QoS的预测.最后,本文基于真实Web服务数据集WSRec进行了对比实验,实验结果表明了本文所提算法的有效性.
In order to improve the QoS prediction efficiency of Web services and provide users with high quality Web services, an extended matrix factorization algorithm based on nearest neighbors is proposed.This method firstly uses the location information of users and services to select the nearest neighbor of users and services, Overcoming the problem of inaccurate or insufficient utilization of the location information of the Web service by the traditional QoS prediction algorithm.Then the neighborhood information is integrated into the matrix decomposition framework to improve the problem of insufficient utilization of the local information by the matrix decomposition technology in the QoS prediction, Algorithm is used to predict the QoS.Finally, a comparison experiment is carried out based on WSRec of real Web service data set, and the experimental results show the effectiveness of the proposed algorithm.