Terry Therneau wrote: > I have a problem whose solution requires non-negative least squares. That is > minimize sum(y - Xbeta)^2 subject to beta >=0 > > Splus has the nnls.fit command. Is there an R alternative? There is a package nnls on CRAN by Kate Mullen and Ivo van Stokkum: http://cran.r-project.org/web/packages/nnls/index.html HTH, Tobias