A method and procedures for automated forecasting of students learning outcomes have been developed, distinguished by the use of heterogeneous factors: the results of taking the Unified State Examination in Russian, mathematics, physics and computer science, assessments of academic performance and parameters of students' cognitive-style potential. The application of the developed method and procedures allows to increase the accuracy of forecasting the results of student learning. Analyzed cases predict the average scores in the interval of the first year and the entire period of study in the university. Developed scenarios for the use of hybrid models of classification and regression for predicting learning outcomes in the disciplines. The prediction procedures are implemented in the OntoMASTER network software complex. The method is designed to improve the accuracy of prediction and the validity of controlling didactic effects (resources) based on dynamically changing data obtained in the learning process.

Authors: E. E. Kotova, A. S. Pisarev

Direction: Informatics, Computer Technologies And Control

Keywords: Predicting student performance, learning process, multiple regression, cognitive potential, intellectual agents

View full article