GAM Procedure
The GAM procedure fits generalized additive models as those models are defined by Hastie and Tibshirani (1990).
This procedure provides powerful tools for nonparametric regression and smoothing.
Nonparametric regression relaxes the usual assumption of linearity and enables you to uncover relationships
between the independent variables and the dependent variable that might otherwise be missed.
The generalized additive models fit by the GAM procedure combine an additivity assumption (Stone 1985) that enables
relatively many nonparametric relationships to be explored simultaneously and the distributional flexibility of generalized
linear models (Nelder and Wedderburn 1972). The following are highlights of the procedure's features:
 permits the following smoothing effects:
 smoothing spline (SPLINE)
 local regression (LOESS)
 bivariate thinplate smoothing spline (SPLINE2)
 supports the following distributions families for the response variables:
 gaussian (continuous response variables)
 binomial (binary response variables)
 Poisson (nonnegative discrete response variables)
 gamma (positive continuous response variables)
 inverse gaussian (positive continuous response variables)

 supports the use of multidimensional data
 fits both generalized semiparametric additive models and generalized additive models
 enables you to choose a particular model by specifying the model degrees of freedom or smoothing parameter
 performs BY group processing, which enables you to obtain separate analyses on grouped observations
 scores new data sets
 creates an output data set that contains diagnostic measures
 creates a SAS data set that corresponds to any output table
 automatically creates graphs by using ODS Graphics

For further details see the GAM Procedure
Examples