中文摘要：

近年来社会安全事件频繁发生,给人民群众的生命和财产带来了严重损害。文章基于大规模时序数据,通过挖掘事件触发因素,利用多维时间序列模型量化分析其与社会安全事件发生的关联关系,并对未来事件的发生数量进行预测。另外,提出一种基于态势主导的多维时间序列相似性度量方法,量化分析不同类别事件之间发展趋势的相似程度,并对三类具体的社会安全事件进行相关分析及预测。实验表明,从时序数据角度分析可以很好地挖掘触发事件的隐形因素,并较为准确地估计事件发生数目和事件发展趋势,为管理者预防和控制此类事件的发生提供了一种新的思路和方法。

英文摘要：

In recent years the frequentoccurring of social security events has led serious damage to masses＇ life and property security. Based on large-scale time series data,this paper quantitatively analyzes the correlation between the trigger factors and the happening of social security events,then predicts the number of security events that may happen in the future. In addition,this paper presents a multi-dimensional time series similarity measurement method which is based on situational dominant,trying to quantitatively analyze the similarity of development tendency among different kinds of events,and make correlation analysis and predictiontowards three kinds of specific social security events. The experiment result shows that time series analysis can well mine the invisible trigger factorsand accurately estimate the number and tendency of public security events＇ happening. It can provide a new thought and method for administrators to prevent and control the happening of these kinds of events.