A Machine Learning-Based Botnet Malicious Domain Detection Technique for New Business  

作　　者：Aohan Mei Zekun Chen Jing Zhao Dequan Yang 

机构地区：[1]School of Cyberspace Science and Technology,Beijing Institute of Technology,Beijing 100081,China [2]School of Continuing Education,Beijing Institute of Technology,Beijing 100081,China [3]Network Information Technology Center,Beijing Institute of Technology,Beijing 100081,China

出　　处：《国际计算机前沿大会会议论文集》2023年第2期191-201,共11页International Conference of Pioneering Computer Scientists, Engineers and Educators（ICPCSEE）

基　　金：Supported by Hainan Provincial National Science Foundation of China,621MS0789.

摘　　要：In the new network business,the danger of botnets should not be underestimated.Botnets often generatemalicious domain names through DGAs to enable communication with command and control servers(C&C)and then receive commands from the botmaster,carrying out further attack activities.Therefore,a system based onmachine learning to dichotomizeDNSdomain access is designed,which can instantly detectDGAdomain names and thus quickly dispose of infected computers to avoid spreading the virus and further damage.In the comparison,the bidirectional LSTM model slightly outperformed the unidirectional LSTM network and achieved 99%accuracy in the open dataset classification task.

关 键 词：BOTNET Machine Learning LSTM Domain Generation Algorithm Detection 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]