移动云环境下基于预测图的动态协调过程推荐  

Dynamic Coordination Process based on Predictive Graph in Mobile Cloud Environment

作　　者：侯小毛 张福泉[2] 刘定 HOU Xiao-mao;ZHANG Fu-quan;LIU Ding(School of electronic information,Hunan Institute of information,Changsha 410105,China;School of Software,Beijing Institute of Technology,Beijin 100081,China)

机构地区：[1]湖南信息学院电子信息学院,湖南长沙410105 [2]北京理工大学软件学院,北京100081

出　　处：《聊城大学学报（自然科学版）》2017年第4期96-100,106,共6页Journal of Liaocheng University:Natural Science Edition

基　　金：湖南省教育科学规划课题项目(XJK17CGD027)资助

摘　　要：为了提高移动云计算环境下服务推荐的准确度和效率,提出了将预测图与动态协调相结合的方法.预测图优化模式采用三种策略完成顶点选择,并将顶点移出顺序作为预测顺序.在动态协调推荐策略设计中,先将客户、服务和属性三者进行随机变换,接着进行预测图近邻矩阵化处理和相似度模拟,最后完成近邻计算,从而为云计算用户提出个性化服务推荐.经过实验证明,相较于传统迭代式的预测推荐方法,本文方法在预测准确度方面精度更高,时间复杂度更低.In order to improve the accuracy and efficiency of service recommendation in mobile cloud computing environment,a method of combining predictive graph with dynamic coordination was proposed.In the predictive graph optimization model,three strategies were adopted to complete the vertex selection,and the order of the vertices removal was taken as the prediction order.In the dynamic coordination recommendation strategy design,firstly,random transformed customer,service and attribute,then worked out predictive graph the nearest neighbor matrix processing and similarity simulation,finally completed the nearest neighbor calculation,thus put forward the user personalized service recommendation for cloud computing.Experiments had proved that compared with the traditional iterative prediction recommendation method,this method had higher accuracy in prediction accuracy and lower time complexity.

关 键 词：移动云计算 预测图 动态协调推荐 近邻矩阵 相似度 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]