An Efficient Malware Detection and Classification System Using Deep AutoEncoding Technique

Malware detection became a perilous challenge in cybersecurity due to the continuous evolution of malicious software and increasing attack vectors. This investigation delivers a malware detection framework which combines the power of Deep AutoEncoder and random forest. 58,596 specimens to take advantage of plausible pre-processed datasets, ensuring robust feature representation and classification accuracy stacked ensemble models designed to provide predictive performance has been enhanced, providing the most comprehensive analysis method compared to traditional single-model methods. Through extensive tests and cross-validation, the proposed algorithm exhibits good performance in key parameters such as accuracy. The results establish how hybrid ensemble learning methods serve to improve the malware detection, especially the exposure of high- dimensional data and complex feature interactions. This research contributes to improving cybersecurity by providing effective and scalable solutions to detect the sophisticated malware threats.