Shared Software

FUNFITS -Curves and Surfaces

FUNFITS is a collection of programs written in S, C and FORTRAN for curve and function fitting. The major methods implemented as S functions include neural net regression, thin plate spline regression, spatial process estimation (kriging), nonlinear autoregressive time series models, splint interpolation and spline smoothing, quantile spline regression, normal kernel density estimation and regression and space filling design estimation.

Graff - Design of Experiments

Design of experiments and graphical analysis of experimental designs including fractional factorial, response surface and robust designs.

GLIMMIX - Generalized Linear Mixed Models

This SAS macro uses iteratively reweighted likelihoods to fit a generalized linear mixed model (see Wolfinger and O'Connell, 1993, "Generalized Linear Mixed Models: A Pseudo-Likelihood Approach," Journal of Statistical Computation and Simulation, 48, p. 233-243). By default, %GLIMMIX uses maximum likelihood to find the parameter estimates if there are no random components and restricted/residual maximum likelihood (REML) if there are. The macro calls PROC MIXED iteratively until convergence, which is decided using the relative deviation of the parameter estimates or the estimated covariance matrix, depending upon if there are no random components. An extra-dispersion scale parameter is estimated by default.

Calib - Calibration and Assay Development

S-PLUS functions for calibration and assay development using the four parameter logistic and other calibration models. The functions are particularly useful for immunoassay (RIA) and enzyme linked immunosorbent assay (ELISA) applications.

Spatial Regression - Abstract and SAS Code

This SAS code describes and fits spatial regression models in multi-dimensional design space (see O'Connell and Wolfinger, 1997, "Spatial Regression Models, Response Surfaces, and Process Optimization", Journal of Computational and Graphical Statistics, Vol. 6, No. 2, pp. 224-241). Predicted surfaces are driven by the covariance structures of the models. Several structures, isotropic and anisotropic, are considered and connections with thin plate splines are reviewed. Estimation of covariance parameters is achieved via maximum likelihood and residual maximum likelihood. A feature of the spatial regression approach is the visually appealing graphical summaries that are produced. Relevant design issues are briefly discussed and spatial designs, such as the packing designs available in Gosset, are suggested as a suitable design complement.