Application of machine learning algorithms for obtaining the predictive evaluation fuzzy utility of participation of the unemployed in the programs vocational training and retraining

Abstract

Evaluation of the effectiveness of vocational training programs and retraining of the unemployed is quite difficult to investigate the problem. The complexity is connected with the versatility of the concept of efficiency itself, with limited opportunities for obtaining reliable data on the unemployed after passing the program, with the inability to separate the effect of passing the program from the effects of other factors affecting the unemployed's position in the labor market. Social and economic efficiency can be viewed both from the position of the state implementing vocational training and retraining programs on the labor market, and from the position of the unemployed participating in the programs. The article analyzes the possibility of applying machine learning methods to obtain a prognostic estimate of the fuzzy utility of participation of unemployed in training and retraining programs. As a criterion of effectiveness, multi-criteria fuzzy linguistic usefulness is used, which is calculated on the basis of the apparatus of special linguistic lotteries and allows to reveal the subjective usefulness of the program for a specific unemployed person. Machine learning methods, using as a learning set (examples) a multidimensional sample of respondents with a known linguistic utility, teach the machine (program) to determine fuzzy utility for respondents - candidates for participation in programs.