BotDetector: An extreme learning machine‐based Internet of Things botnet detection model

Authors:
Publisher website
Google Scholar
Abstract
The development of artificial intelligence has brought new methods for botnet detection. For better performance, deep learning (DL) is more and more widely employed to botnet detecting. The existing DL‐based botnet detection methods require lots of computing resources and running time. While in the real Internet of Things (IoT) environment, real‐time and low computing consumption are much needed. Therefore, the DL‐based methods seem to be powerless in real‐time IoT scenarios. For these reasons, this article proposes a botnet detection model based on extreme learning machine, named BotDetector, which can directly obtain network stream files and quickly learn without data processing to extract botnet traffic characteristics. Experiments show that BotDetector has a good performance, which can identify botnets accurately with great reduction the time consumption and resource consumption. Furthermore, BotDetector has strong applicability in real IoT scenes.
Funding Information
  • The Foundation of the Education Department of Fujian Province, China (No.JAT190025)
  • The National Natural Science Foundation of China (No.61672159, No.61872091, No.U18042631, No.61702105)
  • the Science Foundation of the Fujian Province, China (No.2018J01793, No.2018J01800)

This publication has 39 references indexed in Scilit: