Intelligent Control System for IT Specialists Training Based on Denotative Analytics

Abstract

The study identifies the growing problem of mismatch between the competencies of IT graduates of educational institutions and the requirements of employers. The goal is to develop a set of models and algorithms, as well as software that implements them to improve the efficiency of IT training. The analysis of existing approaches to its solution is made. Prospective directions of specialist modeling are revealed. The methods of labor market analysis in the field of IT-sphere based on vacancies from aggregator sites, approaches to individualization of educational programs and involvement of employers in the training process, as well as means of automation of specialist modeling were considered. Ontological approaches to specialist modeling based on expert knowledge are also studied. The concept of organizing IT training based on open weakly structured data has been developed. An approach based on natural language processing and denotative analytics is described. The proposed concept includes monitoring and collecting data on vacancies from aggregator sites with a given frequency; their subsequent processing to build a dynamic model of an IT specialist, describing the key concepts necessary for professional activity and satisfying employers' requirements; enriching the resulting model with open data on the subject area and information contained in regulatory documents; building a graduate specialist model, taking into account the links between the key concepts by constructing a denotative graph to reconstruct the hierarchy of the enriched set of concepts; forming on its basis a set of recommendations for drawing up and adjusting educational programs. The model of specialist from employer's point of view based on fuzzy sets, model of specialist-graduate including data enrichment algorithm, as well as the algorithm of recommendations formation for making and correction of educational programs were developed. The software for decision support system that implements the mentioned models and algorithms has been developed. Experimental testing with the involvement of experts from the IT industry was conducted. Positive results were obtained as a result of the experiment. The developed system can be used for drawing up and adjusting work programs of disciplines, professional retraining programs, programs of additional education in the field of directions related to information technology.