Estimation of Polytropic Points of Digital Characteristic Variance in Flat Panel Data

Abstract

Flat panel data model has many advantages, such as increasing the degree of freedom between variables, reducing multicollinearity, obtaining more effective estimators and so on. The change of variance in its data means the change of some risk, so studying the variance change point of flat panel data can effectively control the risk. In this paper, the quasi maximum likelihood estimate and cumulative sum estimate are used to reasonably estimate the variance change point in panel data, and further combined with the binary segmentation method, it is extended to the polytropic point problem. Theoretical reasoning and numerical simulation show that the proposed method is reasonable.