基于改进的K-means聚类的学校恐怖袭击风险评估方法

Risk Assessment Method of School Terrorist Attack Based on Improved K-Means Clustering

基于改进的K-means聚类算法采用已有的恐怖袭击事件,构建不同袭击手段对不同袭击目标的风险进行评估分级的模型。首先,根据评估需求以及事件特点统计恐怖袭击事件样本数据;其次,针对K-means算法的恐怖袭击风险评估模型的缺陷采取改进方法,对聚类初始点的选择方法进行改良;再次,基于上述改进构建恐怖袭击风险评估模型;最后,选取1970~2019年全球发生的4 606起学校恐怖袭击事件进行风险评估,同时使用轮廓系数进行效果评价,以检验改进模型的有效性。结果提出改进的K-means聚类恐怖主义风险评估模型通过计算轮廓系数评价其聚类效果优于改进之前的恐怖主义风险评估模型,所构建的恐怖主义风险评估模型可以在不依赖于专家打分、确定权重的情况下基于已发生的事件得出较为客观的风险评估结果。

Based on the improved k-means clustering algorithm and using the existing terrorist attacks, this paper constructs a model to evaluate and classify the risks of different attack means to different attack targets. Firstly, the sample data of terrorist attacks are counted according to the evaluation needs and event characteristics, Secondly, aiming at the defects of the terrorist attack risk assessment model of K-means algorithm, an improved method is adopted to improve the selection method of clustering initial point, Thirdly, based on the above improvements, build a terrorist attack risk assessment model, Finally, 4 606 Global School terrorist attacks from 1970 to 2019 are selected for risk assessment, and the contour coefficient is used for effect evaluation to test the effectiveness of the improved model. In terms of results, the improved k-means clustering terrorism risk assessment model proposed in this paper is better than the improved terrorism risk assessment model by calculating the contour coefficient. The constructed terrorism risk assessment model can obtain more objective risk assessment results based on the events that have occurred without relying on experts to score and determine the weight.