Answers are grounded in this paper and cite their sources.
Li suggested the method of sliced inverse regression for dimension reduction in regressions with a univariate response. In this article we extend that method to multivariate regressions by introducing a new way of performing the slicing. This method applies for any number of response variables and may be particularly useful at the outset of an analysis, before positing a multivariate model. The emphasis is on application; no new asymptotic theory is presented.
No takes yet. Share an insight, caveat, or question.
Setodji et al. (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: