Selected Economic Time Series Analysis Using the Fuzzy Linear Regression

The adequacy of mathematical models of economic systems is reduced by the complexity of their structures, the number of parameters and influencing factors. The mathematical regression model assumes that the structure and functional dependence of the input and output variables of the modeled system is precisely defined. However, real systems are complex and indeterminate, and their adequate models must formalize their vague phenomenon. Artificial intelligence methods use fuzzy set mathematics and fuzzy logic approaches to synthesize models of indeterminate systems. We provided our research of defined fuzzy linear regression models using data series of economic variables, namely the evolution of the discount rate, inflation rate and the rate of unemployment between 2019 and 2021. These data series were chosen with regard to the selected economic cycle before, during and after the Covid-19 pandemy. It is precisely due to the cyclical development of the economy that some level of uncertainty and vagueness of data of monitored variables is manifested. Results of the work reflect outputs of the proposed fuzzy regression model of indeterminate variables during the selected time series. These confirmed the assumptions of the authors that there is a mutual interdependence between the selected economic variables, in particular the amount of the discount rate in relation to the inflation rate, the amount of the inflation rate in relation to the rate of unemployment and thus the amount of discount rate in relation to the rate. The existence of time lags in deciding on economic policy measures and their subsequent implementation was also confirmed in all cases, even during the analyzed time series of three years. Only variable unemployment behaved less standardly, as its essence in many respects lies outside of purely pure market mechanism and is under the influence of market inelasticity, legal measures, free movement of labor in the EU, etc.