Modeling the impact of University students research work on the results of their final certification

Abstract

The problem of the quality of education is formulated as the central problem of the educational process of the higher education institution. It is emphasized that the final certification is an integral indicator that takes into account all the knowledge and skills acquired during the period of study in various disciplines and other activities, one of which research work of students (NIRS) is. The task of predicting the influence of students' research activities on the results of their final certification is formulated. Methods of linear multifactor regression and artificial neural networks as a possible mathematical toolkit for predicting are described. It is shown that the best predicting result is provided by the method of artificial neural networks with a perceptron architecture with 8 input factors and two hidden layers with 5 neurons in each. It is indicated that the proposed approach to predicting can be applied when planning the department's activities, for example, when correcting the curriculum of specialties, syllabuses of scientific disciplines, while adjusting the department's management strategy regarding the interaction of students with academic supervisors.