Comprehensive Framework for the Detec tion and identification of Android malware using Deep Learning
M. K. Nallakaruppan, Rajesh Kumar Dhanaraj · International Journal on Engineering Artificial Intelligence Management Decision Support and Policies · 2025
Android malware is a great concern to mobile devices since it compromises the security of users and their privacy. Due to the high variability and sophistication of contemporary malicious applications, advanced approaches to detect and rec ognize threats are needed. Before discussing the proposed work, let us look at the related studies Briefly, the following is the summary of this paper. Dynamic and static features of the applications are used, and two deep learning models, namely, a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN), are used for further classification. The models are fine-tuned and trained using a synthetic dataset of Android applications’ features including permissions, API calls, behavior characteristics, etc. Following that we evaluate these two models for their accuracy, precision, recall, and computational time. A compari son of the results shows that the CNN model has better performance in feature extraction while the RNN model has done better in modeling sequence patterns. The combination of deep learning with feature engineering shows promising ben efits in the application of real-time malware identification systems.