Research methods of data mining for processing test results

Abstract

The work is devoted to the study of data mining methods for processing test results using the Python language.
The aim of this work is to study methods of data mining using Python.
To achieve this goal it is necessary to solve the following tasks:
- study of the stages of data mining, features of the preparation and study of data;
- study of data mining problems and methods for solving them;
- research of Python language tools for solving data mining problems;
- a practical solution to the problems of IAD: building data mining models using Python;
- analysis of the quality of the constructed models.
In this work, an intellectual analysis of the data obtained from the Moodle learning management system was carried out.
In the work, data mining problems, methods for solving them, and also Python tools for solving these problems were studied.
As the studied data set, the results of testing students of the ITS department for 2018/2019 at the course "Applied Neuro-Fuzzy Systems" were selected. Students took 2 intermediate tests and one exam test in the Moodle distance learning system.
In this paper, the task of clustering test tasks was posed and solved. This task was solved using the Python language, but the clustering results did not match. Solving the problem using Python tools allowed us to analyze test tasks, which are characterized by the absence of a correlation between the correctness of the answer and other answers to testing questions, and also to identify the relationship of such test tasks for testing conducted.