Efficient Cyber Attack Detection in Wireless Sensor Networks Using ANOVA F-Test and XGBoost

Akshat Gaurav, Brij B. Gupta, Kwok Tai Chui

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

1 Citation (Scopus)

Abstract

Wireless Sensor Networks (WSNs) are indispensable in many different applications. Hence, reliability and security of these systems depend on effective detection systems. In this work, we use XGBoost for classification and the ANOVA F-Test for feature selection to provide a model for cyber threat detection in WSNs. We picked the best features using ANOVA F-Test. Particularly excelling in identifying TDMA and Blackhole attack types, our XGBoost model obtained a 90% accuracy, surpassing Logistic Regression, SVM, Gradient Boosting, and Catboost in recall and F1-score across most attack classes.

Original languageEnglish
Title of host publication2024 IEEE 21st India Council International Conference, INDICON 2024
ISBN (Electronic)9798350391282
DOIs
Publication statusPublished - 2024
Event21st IEEE India Council International Conference, INDICON 2024 - Kharagpur, India
Duration: 19 Dec 202421 Dec 2024

Publication series

Name2024 IEEE 21st India Council International Conference, INDICON 2024

Conference

Conference21st IEEE India Council International Conference, INDICON 2024
Country/TerritoryIndia
CityKharagpur
Period19/12/2421/12/24

Keywords

  • ANOVA F-Test
  • Cyber Attack Detection
  • Feature Selection
  • Wireless Sensor Networks (WSNs)
  • XGBoost

Fingerprint

Dive into the research topics of 'Efficient Cyber Attack Detection in Wireless Sensor Networks Using ANOVA F-Test and XGBoost'. Together they form a unique fingerprint.

Cite this