Abstract:
Malware, short for malicious software, is designed to disrupt the normal operation of a
computer or mobile device, acquire confidential information, and fraudulently gain
access to secure computer networks. With mobile devices increasingly targeted by
sophisticated malware, traditional security methods often fall short. This research
involves preprocessing a comprehensive dataset, employing SMOTE to balance class
distribution, and implementing several machine learning models, including SVM,
Random Forest, and XGBoost. The models were evaluated for accuracy, precision,
recall, and other metrics. The results revealed that SVM, Logistic Regression,
AdaBoost, LightGBM, and XGBoost each achieved an accuracy of 0.95, demonstrating
strong capability in distinguishing between benign and malicious applications. Random
Forest followed closely with an accuracy of 0.94, while K-Nearest Neighbors (KNN)
had the lowest accuracy at 0.91. Ethical considerations such as user privacy, bias
mitigation, and transparency were emphasized, alongside a sustainability plan to reduce
environmental impact. This study demonstrates the effectiveness of machine learning
in mobile security, providing a foundation for further research in optimizing and
expanding these methods.
