IDENTIFICATION AND ANALYSIS OF HETEROSCEDASTICITY AT THE ECONOMETRIC DATA APPROXIMATION

Apply

IDENTIFICATION AND ANALYSIS OF HETEROSCEDASTICITY AT THE ECONOMETRIC DATA APPROXIMATION

Journal Title: Новітні технології - Year 2018, Vol 3, Issue 7

Abstract

The article deals with the analysis of econometric data. In econometrics it is possible to build a large number of fundamentally different concurrent models. The choice of the best model is important scientific and technical problem. The process of building and using an econometric model includes the following steps: building a model specification, obtaining estimates of model parameters, analyzing the effectiveness of the obtained estimates, analyzing the predictive ability of the model. A real example of ecometric data was investigated for the linearity of model. The article represents a new approach to the construction of approximating curves to describe the econometric data at conditions of heteroscedasticity. Two methods were considered: using the elements of cluster analysis and using two-segmented linear regression taking into account the heteroscedasticity index. The second approach is more complex, since it involves finding the optimal abscissa of the switching point. The general equation of polygonal regression was obtained using Heaviside function. The advantages of polygonal regression with heteroscedasticity consideration are validated for solving approximation problems.

Authors and Affiliations

V. Kuzmyn, M. Zaliskyi, O. Kozhokhina

Keywords

PROPOSALS ON IMPROVING METHODS FOR ESTIMATION OF EFFICIENCY OF THE SYSTEM OF CONTROL OF THE TECHNICAL STATE OF RADIO ENGINEERING MEANS TRANSPORT SEA

The article analyzes methods for assessing the effectiveness of monitoring the technical state of radio engineering systems of transport sea, identifies their shortcomings and advantages. It is substantiated that the mos...

EP ID EP422943
DOI 10.31180/2524-0102/2018.3.07.01
Views 144
Downloads 0