一种基于聚类的反恐数据分析方法

A Data Analysis Method for Anti-terrorism Based on Clustering

[目的/意义]分析处理反恐情报有利于增强反恐工作的针对性,提高反恐工作效率。[方法/过程]针对全球恐怖袭击事件数据库(GTD)中的数据格式,提出了一种同时适用于定性、定量变量距离量化的新的距离度量学习模型,并将该模型应用于近邻传播聚类算法,利用GTD数据库验证了算法的有效性,进而分析了2017年发生的尚未有组织宣称负责的恐怖袭击事件。[结果/结论]该算法用于恐怖袭击事件的聚类分析,能够提高锁定恐怖分子的准确性,效果较好。

[Purpose/significance] Analyzing and processing anti-terrorism intelligence is conducive to enhancing the pertinence of anti-terrorism work and improving the efficiency of anti-terrorism work.[Method/process] Aiming at the data format in GTD, the paper proposes a new distance metric learning model which can be used to quantify the distance between qualitative and quantitative variables, and applies the model to affinity propagation clustering algorithm.Also, the paper verifies the effectiveness of algorithm by using GTD database, and then analyzes the terrorist attacks in 2017 that have not yet been claimed responsibility.[Result/conclusion] Appling this algorithm to cluster analysis of terrorist attacks can improve the accuracy of terrorist lockout and achieve better results.