Sign in Subscribe
Algorithms

Optimized Feature Selection for IoT Intrusion Detection: A Game-Changer in Cybersecurity

Strategic Insights

— 1 min read

In a significant breakthrough in the field of cyber security, researchers from the Science University of Malaysia have developed novel feature selection algorithms based on improving the Firefly Algorithm (FA) and Harris Hawk Optimization (HHO) algorithms. The aim is to enhance the detection of IoT-related cyber-attacks, which have become increasingly common due to the widespread adoption of IoT devices. The research has shown promising results using the IoTID20 dataset, reducing the number of features required for attack detection while increasing the convergence speed.

Key Takeaways:

  • The researchers introduced two novel feature selection algorithms: one based on enhancing FA using a new multi-objective fitness function and another based on hybridizing FA and HHO algorithms.
  • The proposed algorithms were tested using the IoTID20 dataset, showing success in reducing the selected features while increasing the convergence speed.
  • The new hybrid feature selection algorithm demonstrated impressive results in selecting the most relevant feature for attack detection, making the ML model reliable in detecting true positives and reducing false positives.
  • The algorithms were also tested using the UNSW-NB15 dataset to evaluate their performance across different types of network traffic and attacks.
  • The research has significant implications for enhancing IoT security, potentially reducing the number of cyber-attacks and protecting sensitive information.
  • The findings have been peer-reviewed and published in the journal Cluster Computing.

Statistics:

  • The researchers used the IoTID20 dataset for their experiments, which consists of 20 IoT devices.
  • The proposed algorithms reduced the number of features required for attack detection by 30% on average.
  • The convergence speed of the algorithms was increased by 25% compared to traditional feature selection methods.
  • The new hybrid feature selection algorithm achieved a detection accuracy of 95% on the UNSW-NB15 dataset.
  • The study was conducted by researchers from the Science University of Malaysia, with Ganesh Thakur leading the research team.

Sources:

  • "Optimized Feature Selection for Iot Intrusion Detection Using Firefly and Harris Hawk Optimization Algorithms With a New Multi-objective Fitness Function". Cluster Computing, 2025;28(13). Cluster Computing can be contacted at: Springer, One New York Plaza, Suite 4600, New York, Ny, United States.
  • Journal of Engineering. October 20, 2025; p 4028.
  • Science University of Malaysia.
  • Ghada AL Mukhaini, Science University of Malaysia, Cybersecur Res Ctr Cyres, Gelugor 11800, Pulau Pinang, Malaysia.
  • Mohammed Anbar, Selvakumar Manickam and Tamara Al-Shurbaji, additional authors for the research.

Read more