REDUCING THE ENTROPY OF EXPERT ASSESSMENTS WHEN USING A COMPLEX QUALITY INDICATOR IN THE FORM OF A NORMALIZED AVERAGE POLYNOMIAL

Subject: the article is devoted to the problem of constructing ratings of homogeneous objects by quality, taking into account several particular qualities, for which it is proposed to use aggregating, idempotent, strictly monotonic, shift-invariant polynomials in accordance with the methodology outlined in the works of Zotyev D.B.: «On normalized averages criteria interpolating expert assessments» and «Normalized averages and the problem of collapsing quality indicators». Goal: construct a quadratic average polynomial and use it to model a general consumer rating of smartphones based on seven-part qualities. When constructing a quadratic average polynomial, it is necessary to assign expert assessments at each iteration according to the method proposed in the work, thereby narrowing the range of expert assessment for all subsequent elements of the test set. Prove the practical applicability of the method for reducing uncertainty in the examination process. Research design: a representative sample was selected for the study and private ratings were constructed based on quality indicators of this test population. By analogy with the expert-statistical method for determining weight coefficients, the coefficients of the average polynomial of degree 2, which approximates expert estimates of complex quality, were found. Results: the practical applicability of the method of reducing uncertainty in the examination process has been proven. A comparison was made of the results obtained using the method of reducing uncertainty in the examination process and the expert-statistical method. An example of constructing a quadratic average polynomial, which significantly increases the number of selected parameters of a complex quality indicator, is considered in detail. A recommendation is proposed for the development of this methodology in order to bring its objectivity to the level of making responsible decisions.

multicriteria optimization, normalized average function, shift-invariant polynomial, aggregating operator, weight coefficient, complex indicator, integral indicator, expert assessment, functional weight, decision making

Makhin A. A. Reducing the entropy of expert assessments when using a complex quality indicator in the form of a normalized average polynomial. Biznes. Obrazovanie. Pravo = Business. Education. Law. 2024;2(67):158—165. DOI: 10.25683/ VOLBI.2024.67.1015.