基于ID3—SMOTE结合算法的社会群体性事件预警模型  

Early Warning Model of Social Group Event Based on ID3-SMOTE Combination Algorithm

作　　者：石拓 魏新蕾 邵旭芬 SHI Tuo;WEI Xin-lei;SHAO Xu-fen(Information Engineering School,Communication University of China,Beijing 100024,China;Zhejiang Yueqing Middle School,Yueqing 325600,China)

机构地区：[1]中国传媒大学信息工程学院,北京100024 [2]浙江省乐清中学,乐清325600

出　　处：《中国传媒大学学报（自然科学版）》2017年第6期9-15,共7页Journal of Communication University of China：Science and Technology

摘　　要：当前国内群体性事件表现出组织化、复杂化、政治化、暴力化的特征,严重影响了社会的和谐稳定。通过科学手段预测群体性事件是预防其发生的有效途径。在以往的群体性事件预警方法中,主要都是通过定性分析或简单的定量分析方法实现预测,相对缺乏科学可靠的数据事实作为支撑。文中笔者通过内部单位获取到近年来发生在我国境内的群体性事件的相关数据,创新性地将机器学习的思路引入群体性事件预警领域,颠覆了针对群体性事件的传统分析方法。从社会科学和自然科学的双重视角出发,我们利用机器学习技术科学预测群体性事件。这对政府在处置群体性事件过程中科学决策、有效预防和快速反应具有重要指导意义。At present,the mass incidents in China show the characteristics of organization,complexity,politics and violence,and seriously affect the social harmony and stability.To predict mass events through scientific means is an effective way to prevent its occurrence.The past group events warning methods were mainly through qualitative analysis or simple quantitative analysis to predict the occurrence of social group events,relatively lack of scientific and reliable data facts as a support.In this paper we obtain relevant group events data occurred in China during recent years through internal units,innovatively introduce machine learning into the field of mass incidents,and get the subversion of the traditional analysis method of group events.From the dual perspectives of social science and natural science,we use machine learning technology to predict mass events scientifically.It has important guiding significance for the government in the process of dealing with mass incidents,scientific decision-making,effective prevention and rapid response.

关 键 词：群体性事件 分类 决策树 ID3 SMOTE 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]