MDIot: IoT device identification method based on traffic characteristics

  • Hanxi Zheng
  • , Ruijun Liu
  • , Huanpu Yin*
  • , Haisheng Li
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In recent years, the rapid growth of the Internet of Things (IoT) devices has brought with it security risks, highlighting the need for management for IoT devices. Identification and classification are key components of devices management. However, existing work on IoT device identification is based on prior knowledge to manually extract features, resulting in too many redundant features and reducing devices identification accuracy. In this paper, we propose amodel called MDIoT. Our model use the multi-votingmethod for feature selection, and is combined with Multiobjective Genetic Algorithm to reduce redundancy features and noise. Experimental results show that our feature selection method is more efficient in feature selection and the classification accuracy is over 95%.

Original languageEnglish
Title of host publicationProceedings of 2023 Chinese Intelligent Systems Conference - Volume II
EditorsYingmin Jia, Weicun Zhang, Yongling Fu, Jiqiang Wang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages231-239
Number of pages9
ISBN (Print)9789819968817
DOIs
StatePublished - 2023
Externally publishedYes
Event19th Chinese Intelligent Systems Conference, CISC 2023 - Ningbo, China
Duration: 14 Oct 202315 Oct 2023

Publication series

NameLecture Notes in Electrical Engineering
Volume1090 LNEE
ISSN (Print)1876-1100
ISSN (Electronic)1876-1119

Conference

Conference19th Chinese Intelligent Systems Conference, CISC 2023
Country/TerritoryChina
CityNingbo
Period14/10/2315/10/23

Keywords

  • Device feature
  • Device identification
  • Internet of Things
  • MDIoT

Fingerprint

Dive into the research topics of 'MDIot: IoT device identification method based on traffic characteristics'. Together they form a unique fingerprint.

Cite this