Applying Ensemble Machine Learning Techniques to Malware Detection

Abstract

Cybersecurity is facing serious problems with the proliferation of malware in the internet world. The ever-changing nature of malicious software makes it difficult for traditional detection technologies to keep up. In order to make malware detection systems more accurate and resilient, this study investigates how to apply ensemble machine learning techniques. Using meta-learning frameworks like stacking and boosting in conjunction with various base models like logistic regression, Gaussian Naïve Bayes, and random forest allows the suggested method to make the most of each model's strengths while reducing their shortcomings. By utilizing a large dataset that includes both malicious and benign samples, the stacking algorithm surpassed the rivals in the prediction process, with a recall, precision, and f1-score of 100 after using the encoding method to convert the dataset from a numerical to a categorical format.