When a Linear Model Just Won’t Do: Fitting Nonlinear Models Using JMP®
Susan Walsh, Technical Support Statistician, SAS
This session will use several data sets to explore the different approaches to fitting nonlinear models in JMP. With the addition of the Fit Curve platform, fitting nonlinear models in JMP has never been easier. The first option addressed will be the use of this platform to examine a series of curves by a grouping variable. Specific attention will be given to comparing the parameter estimates, using analysis of means or the equivalence test. A second set of data will be used to fit a sequence of growth models, comparing the different models to determine which model might be the best for that data. Although the Fit Curve platform includes a collection of popular models, it is always possible that a user will want to fit a model not included in that platform. The paper will use additional data tables to demonstrate the use of the previously existing model library. The final approach addressed will be an example of creating a nonlinear model as a formula in the data table and then fitting it in the nonlinear platform. It will include locking the values of parameters in the formula to compare different models and determine the best fit.