基于公众偏好大数据分析的重大突发事件应急决策方案动态调整方法

Dynamic Adjustment Method of Emergency Decision Scheme for Major Incidents Based on Big Data Analysis of Public Preference

针对在重大突发事件应急过程中如何根据与其相关的公众偏好大数据信息来对事件进行划分以及确定事件风险大小等问题,提出了一种基于公众偏好大数据分析的两阶段聚类算法,将事件现场公众在社交媒体上发布的偏好大数据信息进行聚类分析,识别出多个与事件相关的子事件,并得出每个子事件的客观风险级别。然后,结合专家经验判断,综合得出每个子事件的风险级别,进而选择相对应的方案。在此基础上,根据实际突发事件多阶段演变特点,描述了应急决策中多事件多方案的动态调整过程,考虑方案之间的相关性和不同子事件调整方案对各风险级别子事件的应急处置效果、应对损失以及不同方案之间的转化成本等因素,得出不同情境下的最佳调整方案。通过案例分析说明了该方法的可行性和有效性。

Aiming at the issues of how to divide events and determine the risk of events according to the related public preference information during the process of outsize emergencies,this paper proposes a two-stage clustering algorithm based on public preference big data information.Public preference information about the events published on social media is used for cluster analysis to identify several sub-events related to the event and derive the objective risk level of each sub-event.Then,combined with the experts’experience and judgment,the risk level of each sub-event is comprehensively obtained,and then the corresponding alternative is selected.On this basis,according to the characteristics of the multi-stage evolution of emergency,the dynamic adjustment process of multi-event and multi-altenative in emergency decision-making is described.Considering the correlation between the alternatives,the effects of emergency treatment for sub-events at different levels,and the capacity of dealing with loss,as well as the conversion costs between different programs and other factors,the best adjustment alternative under different scenarios can be obtained.The feasibility and effectiveness of this method are illustrated through case studies.