论文部分内容阅读
In this paper,we establish a novel connection between a commonly used null hypothesis and a rank-reducible varying coefficient model in quantile regression.We use B-spline to approximate the varying coefficients in the rank-reducible model,and reveal that the null hypothesis implies a unidimensional structure of a transformed coefficient matrix of B-spline bases.By evaluating the unidimensional structure,we alleviate the difficulty of testing such hypotheses commonly considered 34 in varying coefficient models.We demonstrate through comprehensive numerical studies that the new method is much more powerful than the rank score test which is widely used in quantile regression literature.