SELF-MODELING REGRESSION FOR LONGITUDINAL DATA: Abstract
Statistical Engineering Division
Seminar Series
Self-Modeling Regression for Longitudinal Data
Naomi S. Altman
Chair, Department of Biostatistics
Cornell University
Self-modleing regression is a semi-parametric method for
modeling curves with similar shape. The shape is modeled
nonparametrically, but the differences among curves is modeled
parametrically. This allows a flexible framework for
modeling families of curves, while retaining the usefulness of
parametric methods for modeling the differences among curves.
The methodology will be demonstrated with the analysis of a
randomized complete block experiment, in which the data are
curves. Use of self-modeling regression for experiments with
response curves will be discussed.