Entropy of expert quality assessments according to multiple criteria
DOI:
https://doi.org/10.17308/meps/2078-9017/2025/10/33-44Keywords:
multi-objective optimization, normalized mean function, shiftinvariant polynomial, aggregation operator, weight coefficient, complex index, integral indicator, expert estimate, smartphones rating, functional weight, decision making, uncertaintyAbstract
Importance: the concept of entropy (according to Shannon) is extended to expert assessments of complex quality indicators of objects used for optimization or ranking. The entropy of expert assessments serves as a measure of their uncertainty, which allows measuring and monitoring its reduction as a result of applying certain methods of reducing uncertainty and, accordingly, increasing the objectivity of fuzzy parameters. Purpose: to introduce the concept of entropy of expert assessments of complex quality and to apply it in the problem of marketing analysis solved using the method of interactive approximation of expert assessments. Research design: based on a comprehensive rating of 10 smartphones by 7 consumer properties: smartphone price in dollar, size of permanent and operational memory in GB, battery capacity in MAH, screen size in CM, quality of the front camera in MP, processor clock frequency in GHZ, a polynomial of degree 2 was constructed and the entropy reduction was calculated when using the interactive approximation method. Results: a practical example related to the task of obtaining a comprehensive rating, i.e., ranking smartphones by several consumer properties, is considered. It is shown that as a result of applying the method of interactive approximation of expert assessments, entropy decreased by 60%.
