上下文感知的移动社交网络服务选择机制研究  

Research on Context-Awareness Mobile SNS Service Selection Mechanism

作　　者：王玉祥[1] 乔秀全[1] 李晓峰[1] 孟洛明[1] 

机构地区：[1]北京邮电大学网络与交换技术国家重点实验室,北京100876

出　　处：《计算机学报》2010年第11期2126-2135,共10页Chinese Journal of Computers

基　　金：国家自然科学基金(60802034;60672122);高等学校博士学科点专项科研基金(20070013026);北京市科技新星计划(2008B50);"新一代宽带无线移动通信网科技重大专项课题(2011ZX03002-002-01)"资助~~

摘　　要：服务选择机制作为为用户提供智能性、主动性应用服务的一项关键支撑技术,是移动社交网络中重要的研究内容之一.文中提出了基于上下文、信任网络和协作过滤算法的移动社交网络服务选择机制,将上下文相似度引入到服务选择的过程中,并且和信任度相结合,构成"用户-服务-上下文"三维协作过滤服务选择模型.该方法提高了服务选择的准确性和可靠性,避免了服务选择的盲目性和随意性.仿真实验表明,与传统的协作过滤推荐算法相比,该服务选择算法具有更高的准确率.The service selection mechanism,as a key enabling technology to provide users with intelligent and proactive services,is one of the important research topics in Mobile SNS（Social Networking Service）.This paper presents the service selection mechanism for mobile SNS based on user context,trust network and collaborative filtering algorithm.By importing the context similarity into the service selection process and also considering users＇ trust value,this paper proposes a ＂user-service-context＂ three-dimensional collaborative filtering model.Based on the user context information,the accuracy and reliability of service selection are improved,and the blindness and arbitrary are also avoided.Evaluation results show that the service selection approach based on context,trust relationships and collaborative filtering algorithm outperforms the traditional collaborative filtering algorithm in the accuracy aspect.

关 键 词：移动社交网络服务 服务选择 上下文感知 协作过滤算法 平均绝对误差(MAE) 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]