Collaborative Computing: Networking, Applications, and Worksharing. 11th International Conference, CollaborateCom 2015, Wuhan, November 10-11, 2015, China. Proceedings

Research Article

Personalized QoS Prediction of Cloud Services via Learning Neighborhood-Based Model

Download
355 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-28910-6_10,
        author={Hao Wu and Jun He and Bo Li and Yijian Pei},
        title={Personalized QoS Prediction of Cloud Services via Learning Neighborhood-Based Model},
        proceedings={Collaborative Computing: Networking, Applications, and Worksharing. 11th International Conference, CollaborateCom 2015, Wuhan, November 10-11, 2015, China. Proceedings},
        proceedings_a={COLLABORATECOM},
        year={2016},
        month={2},
        keywords={Cloud services QoS prediction Neighborhood model Parameter learning},
        doi={10.1007/978-3-319-28910-6_10}
    }
    
  • Hao Wu
    Jun He
    Bo Li
    Yijian Pei
    Year: 2016
    Personalized QoS Prediction of Cloud Services via Learning Neighborhood-Based Model
    COLLABORATECOM
    Springer
    DOI: 10.1007/978-3-319-28910-6_10
Hao Wu1,*, Jun He2,*, Bo Li1,*, Yijian Pei1,*
  • 1: Yunnan University
  • 2: Nanjing University of Information Science and Technology
*Contact email: haowu@ynu.edu.cn, hejun.zz@gmail.com, libo@ynu.edu.cn, pei3p@ynu.edu.cn

Abstract

This paper proposes neighborhood-based approach for QoS-prediction of cloud services by taking advantages of collaborative intelligence. Different from heuristic collaborative filtering and matrix-factorization, we set a formal neighborhood-based prediction framework which allows an efficient global optimization scheme, and then exploits different baseline estimate components to improve predictive performance. To validate our methods, a large-scale QoS-specific dataset which consists of invocation records from 339 service users on 5,825 web services on a world-scale distributed network is used. Experimental results show that the learned neighborhood-based models can overcome existing difficulties of heuristic collaborative filtering methods and achieve superior performance than state-of-the-art prediction methods.