Quantile Regression and Box-Cox Transformation's Logical Integration in Fitting Some Kind of Biostatistical Data

Abstract

Quantile regression can be used for some biostatistical research as the first choice. But direct applications are often hampered by the problem of quantile regression planes' crossing in finite sample cases. After research, the Box-Cox transformation is used to avoid the crossing problem. Integrating Box-Cox transformation with quantile regression, we suggest a new method. And the fitting of salmon lustrousness data is presented to illustrate the use of this method.