This research is for the further development of a new class of multi variate semi-parametric model building methods' known collectively as Smoothing Spline Analysis of Variance, (SS-ANOVA) which are suitable for the analysis of data from large cohort studies, either epidemiologic or clinical trials, with many qualitatively different variables observed over several time points. These methods represent an attempt to obtain flexible empirical relationships between multiple complex responses and predictors. If such models can be fitted to the data, then estimated sensitivities of the responses to various predictors can be obtained and the existence of associations between various variables of interest can be tested. The models reduce to standard parametric models if the data suggest that nonparametric terms in the model are not present. SS-ANOVA models have been built and tested for the prediction of multi variate and multi categorical responses and methods developed which allow the analysis of large complex data sets. The investigators will extend this work in several directions: Development of methods to prescreen large, complex data sets for patterns of relationships that warrant further examination; more sophisticated model selection methods, extension to nonparametric multi variate density estimation for the purpose of uncovering conditional and time dependent relationships among the variables, and the development of threshold models. Data from the Wisconsin Epidemiological Study of Diabetic Retinopathy and the Beaver Dam Eye Study will be used to examine the models under study for their reasonableness and for their ability to answer questions meaningful to the study scientists. The results will have broad applicability to other large epidemiological studies as well as to clinical trials. The research software will be developed into a user friendly form, documented, and made publicly available.
Showing the most recent 10 out of 14 publications