Stochasticity, non-linearity and non-parametricity: statistical challenges in approximating socio-economic processes
Stochasticity, non-linearity and non-parametricity: statistical challenges in approximating socio-economic processes

Summary/Abstract: The aim of this article is to indicate the relevance of the challenges faced by statisticians in the analysis of socio-economic processes, which requires breaking many of the well-trodden paths of analysis. Until now a preference for linear, parametric methods was observed, supposedly allowing the discovery of cause-and-effect relationships. According to Kant’s philosophy, ‘a thing in itself’ is not knowable and from this perspective, the main task of statistics is to approximate reality. In statistical studies, only stochastic approximation is made and determining the stochastic structure of processes is as important as studying their regression. The presented review of the literature leads to the conclusion that stochasticity and nonlinearity are the primary features of socio-economic processes and that nowadays their analysis is most effective when based on non-parametric methods. Thus, the paper presents a basic catalogue of methods used for studying non-linearity, the stochastic character and approximation of economic processes using non-parametric methods. An additional aim of this paper is to emphasize the importance of fundamental statistical monographs to the development of statistical research methodology. The contribution of Polish publications to the advancement of contemporary knowledge is also discussed.

• Details
• Contents