Identifikasi Malware pada Android dengan Algoritma KNN dan Principal Component Analysis

Abstract

Android is the most widely used operating systems for mobile device, making it a main target for malware developers. Malware refers to malicious software designed to compromise user devices, often resulting in data theft, privacy breaches, system exploitation, and unauthorized surveillance. Indonesia, among other countries, has seen a significant rise in such threats. This study aims to enhance malware detection performance on the android platform by employing a dataset from Kaggle titled Android Malware Detection. The proposed method integrates the K-Nearest Neighbors (KNN) algorithm for classification with Principal Component Analysis (PCA) for dimensionality reduction, thereby optimizing processing time. The combined KNN-PCA approach achieved a detection accuracy of 97%. Furthermore, the application of the Synthetic Minority Oversampling Technique (SMOTE) to address class imbalance in the dataset led to further improvements in model performance.