基于RF-Apriori混合算法的关键涉恐特征关联规则挖掘  

Mining Association Rules of Key Terrorism-related Features Based on RF-Apriori Hybrid Algorithm

作　　者：潘翔 郭璇 吴文辉[1] 肖治庭[1] PAN Xiang;GUO Xuan;WU Wenhui;XIAO Zhiting(School of Information and Communication,National Defense University of Science and Technology,Wuhan 430010,China;Armed Police College of CAPF,Chengdu 610213,China)

机构地区：[1]国防科技大学信息通信学院,武汉430010 [2]武警警官学院,成都610213

出　　处：《火力与指挥控制》2022年第7期89-96,共8页Fire Control & Command Control

基　　金：2021年全军军事理论课题基金;四川省人文社会科学重点研究基地反恐怖主义研究中心重点基金资助项目(FK20212D02)。

摘　　要：提出一种基于RF-Apriori混合算法的关键涉恐特征关联规则挖掘方法。通过构建多元特征样本数据集,对样本数据预处理后使用RF算法训练生成涉恐威胁主体识别模型,快速发现关键涉恐特征,使用Aprior算法提取关键涉恐特征的频繁项集及关联规则,对关联规则的有效性、可靠性与关联性进行量化评估。实验表明,基于RF-Apriori混合算法能够有效识别涉恐威胁主体,可合理选择关键涉恐特征,快速提取主要涉恐特征的关联规则,提升反恐情报预警预防效能。A method for mining association rules of key terrorism-related features based on RF-Apriori hybrid algorithm is proposed. By constructing a multi-feature sample data set,after preprocessing the sample data,RF algorithm is used to train and generate a terrorism-related threat subject recognition model,and key terrorism-related features are quickly discovered. Aprior algorithm is used to extract frequent item sets and association rules of key terrorism-related features,and the validity,reliability and relevance of association rules are quantitatively evaluated. The experiments show that the RF-Apriori-Based hybrid algorithm can effectively identify the subject of terrorism-related threats,and can reasonably select the key terrorism-related features,can quickly extract the association rules of the main terrorism-related features,and improve the effectiveness of counter-terrorism intelligence early warning and prevention.

关 键 词：反恐情报 RF算法 APRIORI算法 频繁项集 关联分析 涉恐威胁主体 

分 类 号：G359[文化科学—情报学] G631