[R] ecdf() to nls() - how to transform data?

Jochen1980 info at jochen-bauer.net
Tue Jul 19 02:46:44 CEST 2011


Ok, I think I had a good idea to solve my problem and need someone who second
me on that or tell me what a fool I am :-) .

My problem: I went for curve-fitting with nls() and took knot()-ecdf() for
collecting data for nls-basis-dataframe. I came into trouble, because my
x-y-data of the data.frame() was not "equal". I computed 1000 values and
there were always some ties, those ties were not returned by
knot()-function.

Okay, I thought about it and now I am using 100 percentile-values for the
data-frame - nls() works fine - moreover 100 percentiles are always there,
exceptions won't happen. Is it wise to use 100 cumulative percentile for
curve-fitting-base-data instead of ecdf()-function data?

Thanks in advance.

--
View this message in context: http://r.789695.n4.nabble.com/ecdf-to-nls-how-to-transform-data-tp3671754p3676932.html
Sent from the R help mailing list archive at Nabble.com.



More information about the R-help mailing list