Model of the Dynamics of the State of Educational Content Recommender System

Abstract

The report presents a generalized model of the dynamics of the state of the educational content recommender system. It differs from the known ones by taking into account the factor of changes in user preferences over time and the possibility of adapting to them. The relevance of the report is caused by the need to take into account the dynamic nature of user preferences and changes in interests over time. To solve this problem, the authors propose to apply GERT-networks (Graphical Evaluation and Review Technique). GERT networks allow describing the probability of the system staying in different states over time, which is important for understanding and predicting changes in user interests. As a result of modeling, analytical expressions for calculating the probability-time characteristics of the recommender system dynamics are derived. Experiments were conducted to estimate the probability distribution density function of the formation time of educational content recommendation. The results showed the possibility of estimating the maximum values of the probability distribution density of the task of forming educational content recommendations, and, accordingly, the indicator of the time of forming recommendations.