基于SVM-RFE和粒子群优化算法的恶意域名检测模型  

A SVM-RFE and particle swarm optimization based detection model for malicious domain names

作　　者：赵正利[1] 姜鹏 仲国强 吴建新 ZHAO Zhengli;JIANG Peng;ZHONG Guoqiang;WU Jianxin(Department of Education,Ocean University of China,Qingdao,Shandong 266100,China;Network and Information Center,Ocean University of China,Qingdao,Shandong 266100,China;Faculty of Information Science and Engineering,Ocean University of China,Qingdao,Shandong 266100,China)

机构地区：[1]中国海洋大学教育系,山东青岛266100 [2]中国海洋大学网络与信息中心,山东青岛266100 [3]中国海洋大学信息科学与工程学部,山东青岛266100

出　　处：《福州大学学报（自然科学版）》2023年第5期634-638,共5页Journal of Fuzhou University(Natural Science Edition)

基　　金：中国高校产学研创新基金资助项目(2022IT151)。

摘　　要：本研究利用机器学习和随机搜索算法,提出一种基于SVM-RFE和粒子群优化算法的恶意域名检测模型.分析域名字符特征、解析特征和相关特征,使用SVM-RFE算法进行特征权重排序,通过优化的粒子群算法确定最佳SVM参数和特征选择.实验证明该检测模型具有较好的效率和准确度.This study proposes a domain name detection model based on the SVM-RFE algorithm and particle swarm optimization method using machine learning and stochastic search algorithms.The character characteristics,parsing features and relevant characteristics of the domain name are analyzed.The features aresorted by the SVM-RFE algorithm.The parameters and features of supportvectormachine are automatically determined by the particle swarm optimization algorithm.The experimental results show that the monitoring model has higher detection efficiency and better prediction accuracy.

关 键 词：网络安全 恶意域名 支持向量机 递归特征消除 粒子群算法 

分 类 号：TP302.2[自动化与计算机技术—计算机系统结构]