Predictive Network Intrusion Identification & Mitigation Model Using Deep Learning In E-Learning

Abstract

The world of Information-technology has advanced quickly in recent years and network services have extended throughout all industries.  Internet Technology has changed traditional teaching techniques and developed versatile E-learning models. E-learning  models are a great achievement but are vulnerable to cyber-attacks such as Denial-of-Service (DoS). The aim of the study is to develop a  predictive network intrusion identification and a mitigation model using deep learning in e-learning. The research adopts an anomaly  detection methodology. The research datasets consist of 47,645 instances. These instances were divided into training datasets and test  datasets in the ratio of 80:20 respectively. Deep learning was applied to develop the prediction model. Generative Adversarial Networks  (GANs) and Binary classification was used for augmenting and artificial instance generation. The developed model was able to detect  network intrusion with a prediction accuracy of 99.8%. The results of this study can be applied to respond to the ever-evolving attacks in  e-learning platforms to improve data security and protection.