Detection of Malicious Websites using a three model Ensemble Classifiers

Abstract

Malicious attacks are escalating along with the growth of internet users. As a result of that, it is making malware detection inefficient in the cybersecurity field. There are several Machine Learning Classifiers for the detection of malicious websites. Among them include Support Vector Machine (SVM), Random Forest (RF), and Naïve Bayes (NB) which are popularly used techniques. However, when these classifiers are used as stand-alone classifiers, they still suffer from an accuracy sufficiency issue. As a result of that, a three-ensemble classification model to identify a malicious website attack is proposed in this paper to ensure efficient robust malicious detection. Through this paper, it is feasible to reevaluate the malicious attacks and limit the harm that they can cause in the future. In this paper, Support Vector Machine, Random Forest, and Naïve Bayes were combined to develop an ensemble model for malicious website detection. The performance of the proposed ensemble model was evaluated against the three (3) machine learning classifiers using the same dataset. The results showed that the proposed three-ensemble model is a promising solution for malicious website detection.

Keywords

• ensemble
• machine learning
• support vector machine
• random forest
• naïve bayes
• malicious website