Energy-efficient hybrid detection framework for suspicious software components in mobile applications using echo state networks

This research introduces a novel detection framework for identifying Suspicious Software Components (SSC) in mobile applications, enhancing accuracy and efficiency in cybersecurity. The primary objectives are to develop an advanced threat detection system with improved accuracy, reduce false positive rates, and ensure energy-efficient deployment in resource-constrained environments. The proposed method integrates static feature extraction with deep learning and machine learning (ML) models, leveraging Echo State Networks (ESN) for superior threat classification. Additionally, energy consumption analysis ensures feasibility for deployment in mobile environments. The methodology is validated using the CIC-MalMem-2022 dataset, demonstrating improved detection capabilities while maintaining low false positive rates. Experimental results show that the proposed ESN-based framework achieves an accuracy of 99.85% while maintaining superior energy efficiency. This work advances mobile security by offering a robust, scalable, and energy-efficient solution for malware detection.