项目名称: 基于移动网络挖掘的多维群体行为模型研究
项目编号: No.61202303
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 计算机科学学科
项目作者: 涂来
作者单位: 华中科技大学
项目金额: 23万元
中文摘要: 通过分析大量网络数据,定量化地研究人类社会的群体行为模型,既是人类行为模型理论研究的重要发展方向,也有着重要的实际应用价值。而现有研究在"数据源"上仍多以Internet数据或通联信息为主,群体行为在物理时空中表现出的规律尚未得到充分揭示。本课题拟通过对移动网络数据的挖掘,从群体行为感知、行为建模、预测技术三个方面对群体行为模型展开研究,以建立用户在"时、空、人"三个空间上的多维群体社会行为模型。拟解决其中关键的"群体是否存在,如何感知群体行为?"、"如何描述群体演化规律?"以及"如何预测群体行为?"三个问题。拟采用理论分析结合数据实验的方法,本着从定性假设到定量归纳,从理论演绎到评价反馈的技术路线,展开研究。研究工作有助于揭示在多维物理时空中,群体行为的内在机理,加深人们对自身社会行为规律的理解,有助于引导社会重大事件决策和大型工程规划,具有重要的理论前瞻性和实际应用意义。
中文关键词: 通话数据;社交网络;人类行为模式;城市感知;
英文摘要: Quantitatively analyzing human group behavior by mining petabytes of network data represents the tendency in social behavior research and shows great significance on practical engineering as well. However, most of the existing researches focus on mining the Internet information or the contact collections, which lack in physical information such as locations. Therefore, this proposal aims to build a multi-dimensional group behavior model that includes spatiotemporal and human characteristics by analyzing mobile network data. The research covers group behavior sensing, analyzing and predicting. By mining mobile network data, the research will answer the questions: "how to sense group behaviors?", "how human groups evolve?" and "how to predict group behaviors?". The expected results may help to uncover the nature of group behavior in physical world and society, and help people to understand the rules of human community's evolution. The results may also be applicable to instruct decisions in complex events, large systems design, urban plan and etc., which are valuable in both theory and practical applications.
英文关键词: Call Detail Records;Social Network;Human Mobility Patterns;Urban Sensing;