基于移动终端日志数据的人群特征可视化

发布时间：2018-05-01 22:23

本文选题：数据可视化 + 主题模型　；参考：《软件学报》2016年05期

【摘要】：随着我国移动互联网的迅猛发展,如何从海量移动终端日志数据中提取出有效信息,并进行合理、清晰的可视化分析,为工业界等提供有价值的统计分析功能显得尤为重要.目前,对于移动终端日志数据的研究和分析多是基于对单一属性的统计结果分析,如应用下载排行、用户留存率等.为了进一步挖掘移动终端日志数据背后深层次的隐含信息,更加准确地概括出移动终端用户的特征,提出了一种基于移动应用程序日志数据的人群特征分析与画像计算方法,构造了基于移动应用程序数据的主题模型,并将移动设备用户按照与不同应用主题的相关度进行聚类,得到了具有不同特征的人群,从而提出了基于层次气泡图和Voronoi Treemap的可视化展现与分析方案.进一步将人群特征与时间信息、地理位置信息相结合,从多角度可视化展现人群特征.最后,根据该研究内容,实现了B/S架构的日志数据可视化分析原型系统,并通过案例分析验证了该方法的有效性.
[Abstract]:With the rapid development of mobile Internet in China, how to extract effective information from the massive mobile terminal log data and make reasonable and clear visual analysis is particularly important to provide valuable statistical analysis function for industry and so on. At present, the research and analysis of mobile terminal log data is based on the statistical analysis of a single attribute, such as application download ranking, user retention rate and so on. In order to further excavate the deep hidden information behind the log data of mobile terminal and generalize the features of mobile terminal users more accurately, a method of crowd feature analysis and portrait calculation based on log data of mobile application is proposed. A topic model based on mobile application data is constructed, and the users of mobile devices are clustered according to the relevance of different application topics, and the population with different characteristics is obtained. A visual display and analysis scheme based on hierarchical bubble graph and Voronoi Treemap is proposed. Furthermore, the crowd features are visualized from multiple angles by combining the crowd characteristics with time information and geographical location information. Finally, according to the research content, the prototype system of log data visualization analysis based on B / S architecture is implemented, and the effectiveness of the method is verified by a case study.
【作者单位】： CAD&CG国家重点实验室(浙江大学);浙江工业大学计算机科学与技术学院;
【基金】：国家自然科学基金(61232011) 浙江省自然科学基金(LZ12F02002,LY14F020021) 国家科技支撑计划(2014BAH23 F03)~~
【分类号】：TP391.41

【相似文献】