Educational Data Mining: Current Problems and Solutions

摘要

The generalization and systematization of various scientific views of both domestic and foreign researchers on the problem of educational data mining (EDM) and their application to support decision-making on the educational process management. Particular attention is paid to the problems of automatic extraction of interpreted knowledge from educational data in order to transfer it to users and facilitate the processes of interactive computing and communication among users. The proposed methods and models allow increasing the effectiveness of decision-making process at all stages of the educational process management. The novelty of the work results is a new approach to the extraction of interpretable knowledge based on the use of a special class of neural network models, as well as the hybridization of several approaches developed as part of data mining into a single learning environment in order to implement an end-to-end learning strategy, including the steps of data collecting and analyzing, knowledge extracting and management. The paper presents the fragments of the developed software that provides EDM by individual methods.