摘要
移动智能终端的快速发展为用户的位置服务提供了新的应用,以用户位置行为分析为核心的服务技术具有重要的商业应用价值。用户位置及其活动特点和趋势与其所在位置的实际状况及本人意愿密切联系,用户所在位置的资源和状况信息直接影响了用户的位置行为。本文引入小世界网络模型分析用户的位置行为特征,发现用户基于位置的行为属性和聚类。采用推荐度计算方法描述结点之间的相似性,通过将用户位置作为一个树根,把位置资源作为用户的兴趣结点,将兴趣搜索转换为最短路径计算问题。通过改进的最短路径算法计算根结点到各个结点的推荐度,分析用户最感兴趣的位置资源结点。实验结果表明,采用该方法建立的用户位置行为兴趣模型能够很好地描述用户基于位置的兴趣和意愿,算法在结果精度和计算时间上都具有良好的性能。
The paper introduces a small-world network model to analyze user behavior characteristics and find the behavior of the user attributes and location-based clustering.The method using the recommended calculation method described the similarity between nodes and look at the user location as a root.It looks at the location of the user＇s interest in resources as nodes and convert interests computation problem to search shortest path.An improved algorithm is proposed to calculate the shortest path from the root each node and analyze the most interested node resources of users by his positions.Experimental results show that this method established by the location of the user interest model can describe the location of the user＇s interest and willingness.The algorithm has a good performance in accuracy and computing time.
出处
《电信科学》
北大核心
2012年第2期115-118,共4页
Telecommunications Science
基金
国家自然科学基金资助项目(No.71071145)
浙江省自然科学基金资助项目(No.Y6090027)
宁波市自然科学基金资助项目(No.2009A610076)
关键词
小世界网络
位置服务
用户位置行为分析
最短路径算法
small-world network
location based service
user behavior analysis
shortest path algorithm
