Transportation Statistics and Microsimulation

330    ◾    Appendix: Soft Modeling and Nonparametric Model Building

e JMP soware uses smoothing splines to estimate regression curves

using a so modeling approach. e JMP soware allows the user a trial-

and-error approach using a slider. For all choices of stiness asymptoti-

cally (related to kernel bandwidth) the user gets an R-square statistic and

a sum of squares due to the error for the smoothed t.

Another popular smoothing method is loess; see Cleveland (1979) and

Cleveland and Devlin (1988). is approach is very useful when modeling

nonlinear functions. e choice of smoothing parameter is an important

topic. It might be wise for s