Improved Fitting of Constrained Multivariate Regression Models using Automatic Differentiation

Improved Fitting of Constrained Multivariate Regression Models using Automatic Differentiation

Trevor J Ringrose & Shaun A Forth

Published in
Proceedings of the 15th Symposium on Computational Statistics . Eds. Hardle W and Ronz B. Physica-Verlag, Heidelberg. pp 383--388. Berlin, Germany, 2002

Abstract
Regression models for multivariate response surfaces are proposed, and an illustration is given of numerical maximisation of likelihoods using automatic differentiation being superior to that using finite differences.

Download
PDF: tjr_cs02.pdf

AMOR home