基于用户位置与反向预测的QoS预测方法

摘要

在为用户选择并组合其满意的Web服务时,预测Web服务缺失的服务质量(quality of service,QoS)值是必要的.为解决该问题,提出一种R-SRec模型,加入用户地理位置信息和服务的反向预测,提高预测精准度.在基于用户的协同过滤算法中融入用户地理位置信息,提高参与预测数据的空间相关度;在基于服务的协同过滤算法中加入Web服务反向预测,缓解数据稀疏问题;根据不同的置信度融合两种算法,对缺失的QoS值进行预测.使用真实的数据集与其它3类常用的算法进行比较,实验结果表明,该方法的预测结果精确度更高.%It is necessary to predict the missing quality of service (QoS)value when selecting and combining Web services to fulfill the users.To solve the problem,a R-SRec model was proposed,the location information of the users and the backward prediction of services were added to improve the precision of the prediction.The location information was integrated into the collaborative filtering algorithm based on users'information.In this way,the space relativity of the data involves in prediction was increased.The backward prediction was added in the collaborative filtering algorithm based on services'information to relieve the data sparsity problem.These two predicted values were combined according to different confidence levels.Results of experiment with actual dataset demonstrate that the proposed method provides more accurate predicted results than other there common prediction algorithms.