UDC 005.519.8
APPLICATION OF ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING TO ASSESS THE QUALITY OF COLLECTIVE AGREE-MENT ACTS IN THE FIELD OF EDUCATION
N. I. Tsukanova, Ph.D. (in technical sciences), Associate Professor, Department of computational and applied mathematics, RSREU, Ryazan, Russia;
orcid.org/0000-0001-7337-8037, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
V. V. Aleksandrov, PhD (in social sciences), Associate Professor, Department of Automated Control Systems, RSREU, Ryazan, Russia; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
N. V. Golovkin, master student, Department of computational and applied mathematics, RSREU, Ryazan, Russia; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
O. V. Shurygina, master student, Department of computational and applied mathematics, RSREU, Ryazan, Russia; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
The paper discusses the classification of fragments of organization's Collective Agreement (CA). The aim of the work is to solve the problem of classifying the fragment text of the Collective Agreement (CA) to determine which section and issue of labor relations the fragment is devoted to and whether it contains an additional level of labor rights and guarantees provided to employees through the CA, i.e. determine its Quality. This problem is proposed to be solved using neural networks (NN) and machine learning. The article discusses three variants of classifier architecture, which take into account in different ways the relationship between characteristics section, question and quality. Experimental studies of classifier's performance were carried out depending on its architecture, methods of training sample preprocessing, and hyperparameters values. The article presents and discusses the results obtained during the research process.
Key words: : neural networks, text classification, flat classifier, hierarchical classifier, machine learning, sampling data for training, preprocessing of text data, Python language, Keras library