[R] please help on frag polynoms

joerg-burmester@gmx.net joerg-burmester at gmx.net
Thu Jul 10 10:52:47 CEST 2003

hi there,

can anyone help me on the topic of frag polynoms?

i just heard of a friend of mine, that i could build in a functioon called
fragpoly (he was talking of such a function in the 'stata' language) in order
to improve my process of finding an optimal linear model.

instead of trying a vast amount of transformed inputdata to find the best
fit and then step backwards down to e.g. rank 5 in order to get a smooth curve,
i could start right from the beginning with only very few but therefore very
flexible funktions (litle similar to box-plot) which adopt themselve
automaticaly to an optimal fitt and r-squered.

can anyone tell me if such a flexible adoptive function also exists in r+?
(i could not find anything on the r+ pages with these search words).

waiting desperately for help

More information about the R-help mailing list