Hi, I'm wondering if anyone is aware of any R packages that do kernel smoothing with at least 4 predictors. The dataset I'm working with now has 4 predictors, and I can fit a loess smooth, but loess has the disadvantage that it can produce fitted values outside the range of the response-something a kernel smoother can't do. All the packages I've found so far work for 2 or fewer predictors. If anyone's aware of an R package for kernel smoothing that will work with at least 4 predictors, please let me know. Thanks, Paul Louisell ploua@allstate.com ARPC [[alternative HTML version deleted]]