北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2015, Vol. 38 ›› Issue (4): 91-98.doi: 10.13190/j.jbupt.2015.04.019

• 研究报告 • 上一篇    下一篇

客户端上下文感知的Web服务QoS预测方法

马华1,2, 胡志刚1   

  1. 1. 中南大学 软件学院, 长沙 410075;
    2. 湖南涉外经济学院 信息科学与工程学院, 长沙 410205
  • 收稿日期:2014-08-07 出版日期:2015-08-28 发布日期:2015-08-28
  • 作者简介:马华(1979-),男,博士生,副教授,E-mail:hua.ma@csu.edu.cn;胡志刚(1963-),男,教授,博士生导师.
  • 基金资助:

    国家自然科学基金项目(61272148);湖南省科技计划项目(2014FJ3122,2014FJ3040)

Client Context-Aware Prediction of QoS for Web Services

MA Hua1,2, HU Zhi-gang1   

  1. 1. School of Software, Central South University, Changsha 410075, China;
    2. School of Information Science and Engineering, Hunan International Economics University, Changsha, 410205, China
  • Received:2014-08-07 Online:2015-08-28 Published:2015-08-28

摘要:

随着越来越多的相似服务发布到Internet,服务质量(QoS)已成为用户选择Web服务时的关注焦点.识别用户客户端的上下文特征的差异性,有助于为新用户预测Web服务QoS,但现有研究缺乏对影响用户QoS体验的上下文特征的系统分析.提出了一种客户端上下文感知的Web服务QoS预测方法,该方法通过量化分析客户端的地理位置、网络位置、接入设备等上下文特征,应用模糊层次分析法计算历史用户与当前用户的上下文相似度,并以该相似度结果为指导,结合协同过滤技术,以特征加权合成方法预测Web服务的QoS值.通过实验对比和分析可知,该方法能有效解决"新用户问题",并提高Web服务QoS预测的精度.

关键词: Web服务, 服务质量预测, 协同过滤, 客户端上下文

Abstract:

Since the Web services with similar functions are published into Internet, quality of service (QoS) has played an important role in services selection. The identifying client context features between different users is helpful to predict QoS accurately. However, these context features affecting the experience quality of user have not been analyzed systematically in current studies. A client context-aware prediction approach of QoS for Web services was proposed, in which the client context features, including geographical location, network location, and device were analyzed quantitatively. The fuzzy analytic hierarchy process method was applied to calculate context similarity between current user and history users. From that, the similarity weights fusion method was employed to predict the QoS, integrating the collaborative filtering technology. Experiment analysis indicates that this approach can solve the new user problem and improve the accuracy of QoS prediction of Web services effectively.

Key words: Web services, quality of service prediction, collaborative filtering, client context

中图分类号: