search for: conceving

...resumably, this loss function would be least squares, i.e. sum (f(x)-y)^2. The spline transformations would be linear, quadratic or cubic. I know I can solve this problem using some optimization function in R, but I was wondering if anyone knew of a package that already solved this problem. I am conceving the solution as beginning with equidistant knot locations based on the percentiles of x. Then some routine would optimally shift around the knot locations to minimize the loss function. Mike [[alternative HTML version deleted]]