ИСТИНА

The article investigates problems of a priori estimating of the student's outcomes and by using the automated assessment in digital learning platform Mirera. Special attention is paid to reviewing different architectures of the neural networks and selecting the features used for training and validation of the models. The student's outcome estimating is based on the intermediate assessments. The prior estimating could prevent drop-out of the students and apply an adaptive learning approach to adjust the learning path. Additionally, the teacher uses automated estimating results to identify students who need assistance. The authors focus on implementing adaptive learning in the digital learning platform Mirera using neural networks technology.