TECHNOLOGY OF CLUSTERIZATION ON THE BASIS OF ARTIFICIAL NEURAL NETWORKS FOR OPTIMIZATION OF THE INFORMATION BASIS OF TESTS
Journal Title: Вестник Харьковского национального автомобильно-дорожного университета - Year 2018, Vol 83, Issue
Abstract
The use of new information technologies blurs the boundaries between distance learning and education directly within the school. One of the most important and integral components of education is the quality control, assessment of knowledge and skills acquired by students. The testing system allows determining the level and quality of learning in a distance learning system. Test task generators are used to improve the control system. Goal. To control the complexity of a set of tests, it is necessary to determine the function of dependence of the complexity of test tasks on the input characteristics of these tasks. Optimization of the construction of analytical dependence by traditional methods is impossible. It is necessary to solve the problem of self-organization of the information base using neural network technologies. This will allow developing information and analytical support for the database of test tasks. Methodology. The object of the study is a database of tests. The main procedures are streamlining, systematization and classification. A significant number of objective factors affecting the efficiency of its use, create obstacles to the analysis and prediction of trends in the entire database of test items. Kohonen self-organizing network was used for clustering. It implements the principle of learning without a teacher and the result of its functioning is the formation of classes and the assignment of research images to them. The algorithm for creating, learning and modeling is implemented in Matlab. Kohonen’s network training consists in setting up weights in a certain way. Results. The proposed method allows to optimize the procedure for creating and classifying test tasks over time. It is used clustering based on neural network technology. Originality. The advantage of the proposed algorithm is that the neural network without the participation of the teacher determines what information can be extracted without increasing entropy and which factors are secondary in determining the complexity of the test. The number of calculations is significantly reduced. Practical value. Using the obtained base of test items will improve the quality of student learning and more objectively assess the knowledge gained in the learning process. It will facilitate the work of the teacher in preparing the examination tasks.
Authors and Affiliations
М. V. Kostikova, I. V. Skrypina
Keywords
Related Articles
- EP ID EP484391
- DOI 10.30977/BUL.2219-5548.2018.83.0.78
- Views 101
- Downloads 0
RESEARCH OF THE INFLUENCE OF GEOMETRIC PARAMETERS OF THE COMBINED AXIAL TOOL ON THE INITIAL DISPLACEMENT OF THE AXIS
The article considers the influence of geometric parameters of the combined axial tool in the direction of action of the unbalanced radial force. Various states and the place of action of the unbalanced radial force and...
TECHNOLOGY AND EQUIPMENT FOR SOIL PUNCTURE BY USING PULLING POWER OF A «SCREW BLADE–SOIL» PAIR
The experimental desk for conducting the research in natural ground conditions of the working process for soil puncture by using pulling power of a ‘screw blade-soil’ pair is offered. An original measuring system for mea...
STUDY OF GRINDING MODES EFFECT ON REACTIVITY FEATURE OF MINERAL POWDER
The dependences of the degree of crushing of various materials in various grinding mills, the concentration of active centers on the surface of specific materials are considered. The improvement of interaction of mineral...
МЕТОД ОПРЕДЕЛЕНИЯ ТЕМПЕРАТУРНЫХ НАПРЯЖЕНИЙ В АСФАЛЬТОБЕТОНЕ С ПОМОЩЬЮ МЕТОДА КОНЕЧНЫХ ЭЛЕМЕНТОВ
Предложен метод определения температурных напряжений в асфальтобетоне на основе конечноэлементной модели. Показана возможность быстрой оценки температурных напряжений в асфальтобетоне с возможностью прогнозирования низко...
[b]RESEARCH INTO DURABILITY CHARACTERISTICS OF CONVENTIONAL AND COMPOSITE BULLDOZER IMPLEMENTS[/b]
Theoretical research into bearing capacity of conventional and composite implements has been considered in the paper. Methods to increase durability characteristics of both available and newly developed innovative-geomet...