Introduction
The exponential growth of internet usage has unfortunately paved the way for the expansion of malicious activities online. Among these threats, malicious websites stand out as a significant risk to both individuals and corporations. To combat this, a new robust and efficient model for the detection of various types of malicious websites has been developed, achieving high accuracy.
The Proposed Approach
This innovative model employs a two-step process to enhance detection accuracy:
- Feature Selection with Firefly Algorithm: Initially, a feature selection method based on the Firefly algorithm is used to identify the most relevant features for detecting malicious websites. This step ensures that the model focuses on the most important characteristics, improving its efficiency and accuracy.
- Classification with PSO-Optimized XGBoost: Following feature selection, an optimized version of the XGBoost algorithm is applied to classify websites based on the selected features. The parameters of XGBoost are fine-tuned using the Particle Swarm Optimization (PSO) algorithm, further enhancing the model’s performance.
Model Evaluation
The effectiveness of the model was tested against several benchmark classification algorithms using a dataset of over 36,000 websites. The results were outstanding:
- Binary Classification: The model achieved a remarkable 98.42% classification accuracy and an F1 score of 0.984, surpassing other benchmark methods.
- Multiclass Classification: It consistently maintained over 98% accuracy across each class, demonstrating its robustness and reliability.
Key Findings
The proposed model not only demonstrates exceptional classification accuracy but also maintains high precision and minimal false error rates. This makes it a powerful tool for detecting various types of malicious websites, significantly enhancing cybersecurity measures.
Authors
- Sheikhi Saeid
- Panos Kostakos
Publication Details
- Publication Type: A1 Journal article (peer-reviewed)
- Keywords: Cyber Security, Malicious websites, Malicious websites detection, PSO algorithm, XGBoost
- Published: July 3, 2024
- Full Citation: Sheikhi, S., & Kostakos, P. (2024). Safeguarding cyberspace: Enhancing malicious website detection with PSO-optimized XGBoost and firefly-based feature selection. In Computers & Security (Vol. 142, p. 103885). Elsevier BV. DOI: 10.1016/j.cose.2024.103885
Further Information
For those interested in a more in-depth understanding, the full publication can be accessed here.
About IDUNN Project
The IDUNN project continuously strives to enhance cybersecurity through innovative solutions. By integrating advanced algorithms and machine learning models, we aim to provide robust defenses against emerging cyber threats.
Stay tuned for more updates and advancements from the IDUNN project. If you have any questions or would like more information about our work, feel free to get in touch!