We show that the test statistic for dimension in q-based principal Hessian directions (pHd) is distributed as a linear combination of X2 random variables. Simpler distributions can result depending on the distribution of the predictors and the adequacy of the quadratic model.
Bibliographical noteFunding Information:
This work was supported in part by National Science Foundation Grant DMS-9703777.
Copyright 2008 Elsevier B.V., All rights reserved.
- Asymptotic distribution
- Central subspaces
- Dimension-reduction subspace
- Regression graphics