Optimasi Metode Supervised Learning Dengan Menggunakan Particle Swarm Optimization Untuk Deteksi Malware

Abstract

The purpose of this research is for malware detection to solve problems that arise when users access the internet and download files that have been infiltrated by malware. One of the popular solutions today is to use machine learning techniques to train many malware models by considering special features that allow prediction of whether particular software is malware or harmless using machine learning algorithms. The dataset used is a malware detection dataset from Kaggle, which will then be classified using the ensemble classifier algorithm which belongs to the supervised learning category algorithm. Improve classification with feature optimization using Particle Swarm Optimization (PSO). This study resulted in an accuracy value generated by the Ensemble algorithm of 92%, AUC 0.94%. Then, the classification was optimized with PSO, resulting in an accuracy value increased by 7.32% to 100% accuracy while AUC increased by 0.059 to AUC of 1. From the results of the research produced, feature selection is recommended before building a classification model for malware detection.