欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
随着面向服务计算技术的快速发展,越来越多具有相同或相似功能的Web服务被部署在网络上。用户进行服务选择之前,通常需要根据历史调用信息对未使用过的服务QoS进行预测。由于历史调用信息收集过程缺乏有效的监督和约束机制,所采样的QoS信息往往容易受到结构化噪声污染,从而导致现有方法预测性能急剧下降。为了克服这个困难,通过将Web服务QoS预测问题建模为L2,1范数正则化矩阵补全问题,提出了一类基于结构化噪声矩阵补全的Web服务QoS预测方法。真实数据集上的实验结果表明,该方法不仅能精确地辨识出QoS采样矩阵中噪声行所在位置,而且能对缺失Web服务QoS进行有效预测。
英文摘要：
With the rapid development of service-oriented computing, more and more Web services with the same or similar function are deployed on the Internet. Usually, before selecting the most suitable service, users need to predict QoS of unused services from the service invoking history. Due to the lack of effective supervision and constraint mechanisms, some number of the rows in the QoS sample matrix is often contaminated by the structural noise, which leads to a sharp decrease for QoS prediction performance. In order to address this problem, an efficient Web services QoS prediction approach via matrix completion with structural noise is proposed by formulating Web services QoS prediction problem as a L2,1-norm regularized matrix completion problem. The proposed approach can not only exactly detect the position where the data is contaminated, but also effectively predict the missing QoS values. Finally, experimental results performed on a real public dataset demonstrate the feasibility of our proposed approach.
同期刊论文项目
同项目期刊论文
期刊信息
位置：