摘要
为及时发现城市公共场所人群异常聚集区域,有必要构建突发事件热点区域预测模型。统计分析城市移动基站手机接入量和手机移动轨迹数据,建立人群聚集量阈值计算模型,并利用马尔科夫链构建人群密度预测模型,进而建立突发事件热点区域预测模型。采用该模型对某市庆祝"五一"职工大型文体比赛开、闭幕式期间的人群聚集热点区域进行预测。结果表明,手机聚集与人群聚集之间存在较强的相关关系,依据手机的聚集量能估算出人群的聚集规模;使用马尔科夫链能预测不同区域间的人群转移量和人群聚集密度;用该模型能预测人群可能异常聚集的热点区域,预测精度为92.1%。
To timely discover abnormal crowd assembling regions in public urban places,a high emergency region prediction model should be built through the connecting cell phones amount on mobile base station. With statistic analysis of the connecting amount of urban mobile base station phones and the mobile trajectory data,a threshold calculating model of amount of assembling people was built. And a crowd assembling density prediction model was built based on Markov chain. Then a high emergency region prediction model was built. Based on this model,hot crowd gathering regions during the opening and closing ceremonies of a certain city large sports competition on May Day holiday were predicted. The result shows that there exsits a close correlation between cell phone gathering and crowd gathering,which means gathering amount of cell phones can estimate the gathering crowd scale. And Markov chain also can predict the moving crowd amount and gathering density. With the true dataset,the experiments shows that our prediction accuracy of the abnormal gathering crowd is up to 92. 1%.
出处
《中国安全科学学报》
CAS
CSCD
北大核心
2015年第9期159-164,共6页
China Safety Science Journal
基金
国家自然科学基金资助(61272109)
中央高校基本科研业务费专项资金项目(XDJK2014C110
2042014kf0057)
贵州省科学技术基金资助(黔科合LH字[2014]7538号)
湖北省自然科学基金资助(2014CFB289)
