Efficient Estimation of Interval-Valued Symbolic Data Regression Model

Abstract

In the last two decades, regression analysis with interval-valued type data has received more and more attention. For the interval-valued symbolic data regression, the Minmax method and the center and range (CR) method are two widely used popular estimating approaches. In this paper, to improve the estimating eciency of these two estimating methods, seemingly unrelated regression approach has been applied to take account of the dependence information of two regression models of the Minmax method or the CR method. Finally, real data sets are analysed to examine the performance of our proposed procedure.