Practical aspects of artificial intelligence application in formation of competence model (on the example of railway industry)

Abstract

Aim. To determine the potential of practical application of neural networks as a tool to form a model of competencies of JSC “RZD” employees who provide operation of railway infrastructure objects.Tasks. To estimate the quality of the structure, description and level of detail of the competence model formed by a neural network by means of comparison with traditional sources containing requirements to the competence of a track mounting workman position.Methods. The theoretical (analysis, modeling, specification, classification) and empirical (comparison, description, content analysis) general scientific research methods formed the methodological basis. The generation of the competence model by means of artificial intelligence was carried out with the help of the software complex based on neural networks.Results. The structure and description of the competencies generated by the neural network correspond on the whole to the competency model of a track fitter and unified corporate requirements for the employees of JSC “RZD”. At the same time, the description of some professional competences is insufficient, which indicates the need to form an additional clarifying query to the neural network. The corporate competencies, defined by the neural network, are described in detail and contain more competencies, than the existing model of JSC “RZD” employees’ competencies.Conclusions. In order to prepare a competency model, neural networks serve as an effective tool, which allows to ensure the proper quality of competency structure generation and description. The universality of this approach lies in the possibility to generate a competency model for any position regardless of the industry. The level of inclusiveness of neural networks increases with the development of digital environment, which ensures their accessibility in practice.