A Comparative Analysis for Android Malware Detection Using Machine Learning Models
Harsh Shah, Vidit Shah, Nirmay Soni, Vaishali Vadhavana, Krishna Patel · 2025
This research study explores the usability of machine learning algorithms in classifying whether Android applications are excellent or malicious. The data used here for the study is obtained from Kaggle and named “TUANDROMD.” Both kinds of application samples are present in the data. From application displays and runtime behaviour, pertinent features have been retrieved using methods including K-Nearest Neighbor(KNN), Support Vector Machine(SVM), and Decision Tree classifiers. With regard to accuracy, precision, recall, F1-score, and ROC AUC metrics, every result is thoroughly examined. The provided SVM model achieved the highest possible accuracy and reliability in malware identification. Thus, the research demonstrates how to combine static and dynamic analysis as a fundamental method for detecting malware and future directions regarding improvements in mobile Security, including deep learning and real-time detection systems. The classification accuracy of the classifier model KNN, SVM, and Decision Tree classify a 97.65%, 98.62%, and 99.10% accuracy rate, respectively.