Machine Learning Approach for Malware Detection and Classification using Bio Inspired Algorithms

The prevalence and sophistication of malware today pose a challenge to conventional signature-based malware detection techniques. This paper provides an approach to the identification and classification of malware that makes use of bio-inspired algorithms in conjunction with machine learning (ML) approaches. The main purpose of this paper is to design a system that can accurately recognize and classify malware samples as they are encountered in real time. This paper uses different bio-inspired algorithms, such as Ant Colony Optimization (ACO) and Particle Swarm Optimization (PSO), to achieve this. These algorithms take cues from biological processes. These techniques are used to fine-tune the feature selection process, which plays a major role in determining the quality of the detection and classification models. These methods improve the system’s overall performance by simulating natural evolutionary processes to find the most discriminative and relevant features among a large set of candidate features. To predict the classification accuracy different ML algorithms are used like Logistic Regression, Gaussian Naive Bayes, K Nearest Neighbour, Decision Tree, Multilayer Perceptron, Support Vector Machine, JRip, Random Tree, REPTree along with ACO and PSO. The results shows that bio-inspired algorithms improved the accuracy of the malware detection as compared to the basic ML algorithms.