APPROXIMATION OF FUNCTIONS OF CAUSE-EFFECT RELATIONSHIPS OF PRODUCTION METRICS IN THE ANALYSIS OF DATA IN THE LIVESTOCK INDUSTRY

Abstract

The article highlights the issues of building and analyzing mathematical models obtained by approximation of functions that reflect cause-and-effect relationships of production metrics in the livestock sector of the agricultural economy. The main hypothesis of the study is to determine the significance of the influence on the volume of meat production of such factors as: offspring of calves, offspring of piglets, and average annual milk production. Confirmation of this hypothesis is possible due to the approximate evaluation of the functions that describe the cause-and-effect relationship between the production metrics of the livestock industry. The essence of the causal relationship is identified using a set of metrics on a finite sample of data statistics obtained in the dynamics for 2010–2020. The data sets for the study were previously evaluated by descriptive statistics analysis, which confirmed a small scatter of the data around the expected mean values. The coefficient of variation and correlation were chosen as the key characteristics of the size of the degree of dispersion of values. The sequence of metric influence was assessed based on the significance of parameters approximated by the least squares method for paired and multifactor functions. The algorithm of the proposed evaluation demonstrates the step-by-step process of determining the significance of metric impact and obtaining the final meaningful result. The experiment revealed metrics (calf progeny and average annual milk yield), the importance of which did not correspond to the Student’s t-test assessment and, as a result, did not constitute the essence of the causal effect. The effectiveness of the influence of the piglet offspring factor was confirmed by the evaluation of the significance of the parameters of the approximation of the power function.