Hybrid Artificial Fish Swarm Optimization with Deep Learning-Driven Cloud Assisted Cyberattack Detection
International journal of intelligent engineering and systems · 2024
In the evolving landscape of cloud computing and the Internet of Things (IoT), the Android Operating System (AOS) has emerged as a focal point for cybersecurity efforts, particularly due to its vulnerability to a wide array of cyberattacks.These threats, which include financial loss, privacy breaches, unauthorized access, data integrity compromises, and denial of services (DoS), have accentuated the need for advanced malware detection solutions.This study introduces a pioneering cloud-enabled Hybrid Artificial Fish Swarm Optimization with Deep Learning-Driven Malware Detection (HAFSO-DLMD) technique for Android devices, aiming to enhance the precision of malware identification through deep learning models.The HAFSO-DLMD technique preprocesses the bytecodes of Android applications' classes.dexfiles for input into a Deep Sparse Autoencoder (DSAE) represent a significant innovation in the field.By employing the HAFSO algorithm for optimal hyperparameter tuning, we have demonstrated a substantial improvement in the detection rate of the DSAE model.Our comprehensive experimental evaluation on an Android APK dataset comprising 16,000 samples has underscored the HAFSO-DLMD technique's superior performance, achieving accuracy, precision, recall, and 𝐹 𝑠𝑐𝑜𝑟𝑒 of 99.17%.These results significantly outperform other contemporary approaches, thereby establishing the HAFSO-DLMD method as a potent tool in bolstering Android's cybersecurity infrastructure within cloud environments.