Questions of Managing the Trajectory of E-Learning Using Neural Network Technologies

Abstract

This article discusses the application of neural networks in managing the educational trajectory in electronic educational systems. A model of learning management in an electronic information and educational environment is built, and an algorithm for managing adaptive learning is described. A test example of a neural network is constructed, showing the capabilities and problems of applying neural networks to the tasks of constructing trajectories of students’ adaptive personalized training. The following stages of constructing a neural network model are implemented: determining the structure of a neural network and its software implementation using the PyTorch deep machine learning library; preparing a test training sample for training the neural network; importing the initial data for training the model using the modules of the Pandas library; training the model using torch.optim Adam, Adamax, and Rprop optimizers; visualizing the obtained data in the form of graphs based on the matplotlib library; exporting the obtained numerical results and analysing the data. The paper also states the problems of preparing a training data set for teaching a neural network. The paper concludes that using neural networks in the field of building personalized adaptive educational trajectories has great prospects and will be one of the urgent tasks in the coming years.