Adapting the method of principal Hessian directions as recently proposed by Li (1992), we present a test for aiding an assessment of the adequacy of a regression function.
Bibliographical noteFunding Information:
This work was supported in part by grants from the National Science Foundation. Doug Drake provided help with computing.
- Central dimension-reduction subspace
- Linear combinations of chi-squared variables
- Principal hessian direction
- Regression graphics