基于隐因子分解机的上下文感知的QoS预测算法研究
ontext-Aware QoS Prediction Based on Machine Factorization
QoS即服务质量,指一个网络能够利用各种基础技术,为指定的网络通信提供更好的服务能力。在Web服务的QoS预测中,上下文的信息对于QoS的预测非常关键。传统的QoS预测模型使用协同过滤或者基于聚类的算法,通过对用户和服务进行划分,不断地缩小样本集合。但是这种方法基于用户和服务之间的相似度。在有上下文信息的情况下,维度较大导致存在极大的稀疏性,使得相似度计算不准确。同时,上下文信息中不同维度并非独立作用于QoS结果,通常的相似度计算无法考虑不同维度之间的交叉关系。基于以上的问题,本文提出了一种基于隐因子分解机模型的上下文感知的QoS预测算法。与传统的模型相比,隐因子分解机可以直接将上下文信息进行建模,同时可以将特征之间的高阶交叉关系纳入到模型的计算之中。通过实验结果,可以得出,本文提出的算法在均方根误差和平均绝对误差两项评价指标上,均比之前的算法有了很大的提高。
Quality of Service, QoS is kind of network providing better service capability for network communication with various basic techonologies. Context information is exactly important to QoS prediction. Existing works model QoS prediction with collaborative filtering or matrix Factorization, which can not incorporate context information into the modeling process. Although context-aware QoS prediction with clustering can accurately describe users' requirements in specific context environment, it seldom excavate internal and implicit relationship between users and web services. In this paper, we propose a context-aware QoS prediction model of Web services based on Factorization Machine. Compare to existing works, Factorization Machine models context information directly, and incoporate the cross relationship between features into modeling calculation. Experiment results demonstrate effectiveness of our models.
张成文、胡月
计算技术、计算机技术
隐因子分解机上下文感知QoS预测
Factorization MachineContext-AwareQoS Prediction
张成文,胡月.基于隐因子分解机的上下文感知的QoS预测算法研究[EB/OL].(2016-12-07)[2025-08-03].http://www.paper.edu.cn/releasepaper/content/201612-149.点此复制
评论