Related Experiment Video
Updated: Jun 5, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
468
An optimized ensemble model with advanced feature selection for network intrusion detection
Afaq Ahmed1, Muhammad Asim2, Irshad Ullah1
1School of Computer Science and Engineering, Central South University, Changsha, Hunan, China.
Peerj. Computer Science
|December 9, 2024
Summary
This study introduces Optimized Random Forest (Opt-Forest) for advanced network intrusion detection. Opt-Forest enhances security systems by effectively identifying sophisticated cyber threats, outperforming traditional methods.
Area of Science:
- Computer Science
- Cybersecurity
- Artificial Intelligence
Background:
- Modern networks face evolving cyber threats.
- Traditional intrusion detection systems struggle with sophisticated attacks.
- Need for resilient and adaptive Network Intrusion Detection Systems (NIDS).
Purpose of the Study:
- To develop an enhanced NIDS model for improved threat detection.
- To introduce the Optimized Random Forest (Opt-Forest) ensemble model.
- To increase the adaptability and resilience of NIDS against contemporary cyber threats.
Main Methods:
- Developed the Optimized Random Forest (Opt-Forest) model.
- Integrated genetic algorithms (GAs) for decision forest construction.
- Employed advanced feature selection: Best-First Search, PSO, Evolutionary Search, and Genetic Search.
Main Results:
- Opt-Forest demonstrated superior performance compared to traditional ML models (AbM1, KNN, J48, MLP, SGD, NB, LMT).
- The GA-based approach facilitated wider exploration and avoided local optima for more accurate trees.
- Achieved enhanced accuracy and reduced false alarms in network intrusion detection.
Conclusions:
- The Opt-Forest model significantly enhances NIDS capabilities.
- Genetic algorithms improve decision tree accuracy and compactness.
- Proposed method offers a robust solution for detecting evolving cyber threats.
