Hybrid Deep Learning-Assisted Intrusion Detection Model for Cloud Environment
The aim of this thesis is to improve intrusion detection accuracy and efficiency in cloud by developing and deploying intrusion detection system trained using hybrid deep learning. Hybrid approaches are proving to be successful in protecting the cloud from zero-day cyber-attacks and needs to be explored further specially as majority of the research is done using old benchmark dataset. We ask how intrusions can be accurately detected along with the consideration of the optimal hidden layer selection, gradient-based adaptive weight updation, and optimal trade-off between the error rate and the optimal features using more recent datasets. The research outcome will enrich the traditional intrusion detection models resulting in better overall cloud security for organizations.