The generalize maximum Tsallis entropy estimator in kink regression model

Abstract

© Published under licence by IOP Publishing Ltd. Under the limited information situation, underdetermined or ill-posed problem in statistical inference is likely to arise. To solve these problems the generalized maximum entropy (GME) was proposed. In this study, we apply a generalized maximum Tsallis entropy (Tsallis GME) to estimate the kink regression using Monte Carlo Simulation and find that Tsallis GME performs better than the Least squares and Maximum likelihood estimators when the error is generated from unknown distribution. In addition, we can claim that the GME is a robust estimator and suggest that Tsallis GME can be used as an alternative estimator for kink regression model.