Deep Learning-Based Intrusion Detection for Online Learning Systems
S. Baghavathi Priya, K Sangeetha, V. S. Balaji, TamilSelvi Madeswaran · 2024
Education 4.0 is a domain of intelligent and smart thinking. The main objective of Education 4.0 is to improve the teaching skills and enhance the learning outcomes of the students. Deep learning (DL) algorithms build a model based on sample data, which can be used to predict, identify, and classify patterns in the data. DL algorithms are Convolutional Neural Network (CNN), Multilayer Perceptron, Random Forest, and Support Vector Machine (SVM) that can be incorporated in Education 4.0 to identify intruders and classify malware. It is essential to authenticate the user and data content involved in the learning process through Education 4.0. The CNN is proposed for user authentication based on unique typing patterns. This model recognizes individual-specific patterns within the data, such as identifying users based on their typing behavior. The proposed model provides 5.52 it/s , that is, the rate of iterations, as it has various number of hidden layers. The result “ 0s 114 ms/step ” indicates the time taken to complete one step (or iteration) during the authentication process. The CNN model provides 98.2% of accuracy in predicting the typing feature verification and IP authentication. This will enable the system to detect intruders in the learning process of Education 4.0.