An Approach to Comparative Analysis of Online Programming Education Datasets Based on Machine Learning Algorithms

This paper discusses an approach to performing a comparative analysis of datasets containing information about student activity in different cycles of online programming education. The purpose of the analysis is to identify similarities or differences in student activity both in general based on the results of the educational cycle and on individual tasks offered to students during online training. It is proposed to use such machine learning algorithms as the linear SVM algorithm and the logistic regression algorithm when performing a comparative analysis, assigning different class labels to datasets related to different educational cycles. The results of experimental studies confirming the effectiveness of the proposed approach are presented.