Scientific and Technical Journal

ELECTROTECHNIC AND COMPUTER SYSTEMS

ISSN Print 2221-3937
ISSN Online 2221-3805
USING OF CONVOLUTIONAL NEURAL NETWORKS FOR INFORMATIVE FEATURES DEFINITION INFLUENCING OF THE DISTANCE LEARNING QUALITY
Abstract:

The distance learning quality assessing methods are divided into static and internal. Static methods are based on learning, comparison of the obtained digital data with each other or with the sample, their generalization, interpretation and the formation of scientific and practical conclusions. Internal methods use the idea of analyzing the received digital data using the built-in (or additional) tools of the system. Perspective in our time is the direction with the application of cognitive maps. The cognitive map is a structure of knowledge, a graphic representation of causal relationships between concepts, attributes, indicators, interacting with systems and their blocks. The closest and most accurate implementation of cognitive maps is neural networks.

The purpose of the presented paper is searching of features that have the greatest impact on the education quality. This paper is devoted to using of convolutional neural networks (CNN) in the synthesis of cognitive maps in order to determine information features that affect the level of student’s education of university when using the component of distance learning in the university's educational process. Synthesis of CNN for the construction of cognitive maps occurs as follows: after creating CNN and connecting it to LMS Moodle, data sets from the LMS Moodle database are submitted to the CNN input. When using the filter, the weight of each data set is formed. All data sets are interrelated with a predetermined set of features considered. Based on the results of operations, a characteristic map is formed - those functions whose weight is greatest are selected. Advantages of CNN - high speed of data processing, automation of the process of calculating the weight for each function, high resistance to interference. The most informative features were identified: the number of tests in the course; availability of practical assignment activities; theoretical materials posted on the course; time spent by the student on the course after authorization. Here are the quality features, which can improve the quality of distance learning in general.

Authors:
Keywords
References
  1. Savyuk, L. K. (2004) Legal statistics [Pravovaya statistika]: Textbook, Yurist, Moscow, 588 p.
  2. Moodle DOCS, available at: https://docs.moodle.org/34/en/Main_page
  3. Cognitive Science and Intellectual Tech-nologies (1991) [Kognitivnaya nauka i intellektual'nye tekhnologii]: Ref. Sat. AN SSSR, Institute of Science. inform. on societies. Sciences, Moscow,228 p.
  4. Filonenko, K. M., Fomin, O. O., Pavlenko V. D. (2017) Using of cognitive maps for highlighting the factors of students' success in distance learning [Ispol'zovanie kognitivnykh kart dlya vydeleniya faktorov uspevaemosti studentov distantsionnoi formy obucheniya], Modern Information Technologies 2017 (MIT-2017): Materials of the Seventh International Conference of Students and Young Scientists, ONPU, Odessa, №2, 246 p.
  5. Mandel, M. V. (2013) The method of cognitive analysis for the study of problems of the regional labor market (south of Russia) [Metodika kognitivnogo analiza dlya issledovaniya problem regional'nogo rynka truda (yug Rossii)], Management of economic systems, Southern Federal University, Taganrog
  6. Avdeeva, Z. K., Kovriga, S. V., Makarenko, D. I., Maksimov, V. I. Cognitive approach in management [Kognitivnyi podkhod v upravlenii] (2007), Problems of management, Moscow
  7. Advantages and disadvantages of neural networks as a means for processing knowledge [Dostoinstva i nedostatki neironnykh setei kak sredstva dlya obrabotki znanii], Studopedia, available at: https://studopedia.org/9-161389.html
  8. Garipov, T. I. Tensored neural networks [Tenzorirovannye neironnye seti], M. V. Lomonosov Moscow State University, Moscow, available at: http://www.machinelearning.ru/wiki/images/2/22/2017_417_GaripovTI.pdf
  9. Ostapets, A. Deep Learning, available at: http://www.machinelearning.ru/wiki/images/2/22/Deep.pdf
  10. Dorogoi, Ya. Yu. (2011) Architecture of generalized convolutional neural networks [Arkhitektura obobshchennykh svertochnykh neironnykh setei], News of NTUU "KPI". Informatics, management and computing engineering, Kyiv,              p. 229-234
  11. Understanding Convolutions, available at: http://colah.github.io/posts/2014-07-Understanding-Convolutions/
  12. Fedorova, A. A. (2016) Recognition of English text convolutional neural network [Raspoznavanie angliiskogo teksta svertochnoi neironnoi set'yu], St. Petersburg State University, St. Petersburg, available at: https://dspace.spbu.ru/bitstream/11701/4262/1/VKR.pdf
  13. Druki, A. A. (2014) Application of convolutional neural networks for allocation and recognition of automobile license plates on images with a complex background [Primenenie svertochnykh neironnyi setei dlya vydeleniya i raspoznavaniya avtomobil'nykh nomernykh znakov na izobrazheniyakh so slozhnym fonom], Intellectual systems. Proceedings of Tomsk Polytechnic University, Tomsk, t. 324, № 5, p. 85 - 92
  14. Medvedew A. O., Fomin O. O., Speranskiy V.O. (2017) The diagnostic of retinal diseases by dint of convolutional neural networks, Modern Information Technologies 2017 (MIT-2017): Materials of the Seventh International Conference of Students and Young Scientists, ONPU, Odessa, №2, 246 p.
Published:
Last download:
9 Dec 2018

[ © KarelWintersky ] [ All articles ] [ All authors ]
[ © Odessa National Polytechnic University, 2014-2018. Any use of information from the site is possible only under the condition that the source link! ]