Machine Learning-Driven Malware Detection: Integrating PCA and PSO for Efficient Classification

Due to the exponential growth of digital platforms and the Internet era around us, particularly: cyberattacks facilitated by malware have become increasingly diversified and complex. The increase in malware strains has created an increased demand for advanced detection methods that can handle a large volume of data quickly. In this study, we use machine learning classifiers along with optimum feature vector extraction using principal component analysis (PCA) and Particle swarm optimization (PSO) to get better malware detection and classification. The key classifiers of this method are logistic regression (LR), Naive Bayes (NB), nearest neighbor (KNN), and decision trees (ID3 and C4. 5). PCA and PSO are successful methods for reducing the feature set without hindering or even improving the classification results; therefore, experimental results show that using feature reduction techniques with these classifiers provides (i) very high detection accuracy and (ii) lower computational complexity. These can be seen as evidence for the usefulness of hybrid approaches in enhancing malware detection performance on large-scale data sets.