Algorithm for making recommendations in the electronic educational environment based on Markov stochastic models

This article proposes an algorithm for making recommendations for educational resources in the electronic educational environment. The new approach uses the Markov model for evaluating the content of the systems by ordinary users to form the parameters of the initial state, which characterizes the system new user in the form of assessing the first liked resources (system content) to recommend the system interesting elements to an active user. Thus, the problem of “a cold start” is solved for a new user at the first stages of interaction with the system. This problem is inherent to the system under development, since the e-learning system provides a module for making recommendations, which allows it to be classified as a recommendation-based automated system. The new approach proposes combining the use of the Markov process and the time factor to apply them as a single data source for making recommendations. This approach will be based on the principle of analyzing the access of similar users of the system (the similarity is determined by comparing their profiles) at the same time periods. Ease of use is also an integral part of the system being created. Therefore, at the design stage, it is necessary to think over the ergonomics of the recommendations in the educational system.