基于UGC大数据挖掘的大群体两阶段风险性应急决策方法

针对在重大突发事件应急决策大数据环境下决策者偏好的不确定性及偏离群体一致性导致的风险,提出一种基于UGC大数据挖掘的大群体两阶段风险性应急决策方法。首先,通过数据挖掘和自然语言处理方法从UGC中获取公众对事件的偏好信息并构建应急决策属性体系,利用TF-IDF方法结合专家评估信息确定属性权重;其次,建立一个意见开放式的两阶段决策流程,提出依据决策者意见的可靠度和准确度量化决策风险,利用聚类方法得到相应的成员权重,并使用TOPSIS法对决策方案进行排序。最后通过天津港“8·12”重大爆炸事故的案例分析和对比验证了所提出方法的可行性和有效性。

A Method of Two-stage Risky Emergency Decision for Large Group Based on the UGC Big Data Mining

Aimed at the risk of both the uncertainty and deviation from group consistency of decision makers’ preference in big data environment of the major emergencies decision-making, a method of two-stage risky emergency decision for large group based on the UGC big data mining is proposed in this paper. First, public preference information about events is obtained from UGC through data mining and natural language processing, and the emergency decision attribute system is constructed through TF-IDF method which is combined with expert evaluation information to determine attribute weights; Secondly, an open two-stage decision-making process is put forward to quantify the decision-making risk according to the reliability and accuracy of decision makers’ opinion, using the clustering method to get the corresponding membership weight and using TOPSIS method to rank alternatives. Finally, the case analysis and comparison of the “8 · 12”major explosion in Tianjin Port verify the feasibility and effectiveness of the proposed method.