Оценка значимости признаков на основе нейронных сетей в задачах анализа качества дистанционного обучения

• Дата: 2018
• Автори: Филоненко, Екатерина Михайловна, Рубан, Александр Дмитрович, Фомин, Александр Алексеевич
• Формат: Стаття
• Мова: Russian
• Опубліковано: Kamianets-Podilskyi National Ivan Ohiienko University 2018
• Онлайн доступ: http://mcm-tech.kpnu.edu.ua/article/view/158725
• Назва журналу: Mathematical and computer modelling. Series: Technical sciences

Informatization of modern education contributes to the creation of new methods for the development of educational courses, which significantly reduces the quality of students' education. In this paper, a method for assessing the significance of features when analyzing the quality of the introduction of distance learning in higher education institutions using a neural network is proposed.An algorithm for carrying out a study of significance of features is presented, consisting of three stages: data preparation, neural network modeling and analysis and interpretation of the results of the study.As a training sample, the real data of the university students' training from the Moodle distance learning system was used. This system is actively used as a tool for conducting the educational process at the Odessa National Polytechnic University.Neural network modeling consists in investigating the informative character of the traits after training the neural network. The inputs of neurons were data on the progress of students in the courses, as outputs — their resulting estimate for the course.The values of the matrix of the weights are visualized with the help of graphs and histograms and enable us to analyze the results of the study and clearly confirm the significance of the signs.Thus, the task of assessing the significance of characteristics in the analysis of student learning data in the Moodle distance learning system was solved. The assumption that the weights of the features characterize the level of significance of each investigated feature is substantiated. The most significant features that affect the quality of the introduction of distance learning are highlighted.