Threshold regression for modeling symbolic interval data

Abstract

© Serials Publications Pvt.Ltd. This paper proposes a threshold regression model for symbolic interval data to explain the structural change in interval time series data. In this study, the center method and center-range method are applied in the thresold regression model and the parameters are estimated by Bayesian approach. The study employs Gibb sampler and Metropolis-Hastings alogorisms for drawing the parameter values. The results of simulation study and applications to real data are used to evaluate and show the usefullness of our proposed model.