Re: regression

2009-06-05 Thread kyle smith
I have generalized the code to allow more than one independent variable. The calling convention has changed: instead of separate x and y seqs, there is one "xylist". Each element of xylist is a tuple containing the dependent variable *first*, and then any independent variables. I haven't had ti

Re: regression

2009-05-27 Thread Daniel Lyons
On May 26, 2009, at 10:56 PM, kyle smith wrote: > Thanks for the feedback Daniel, I've incorporated your ideas and re- > uploaded. I'm not sure where you're seeing mutable data structures. I'm hallucinating, that's where. :) > Anyhow, I now only call eval once each time scorer is called, whic

Re: regression

2009-05-26 Thread kyle smith
Thanks for the feedback Daniel, I've incorporated your ideas and re- uploaded. I'm not sure where you're seeing mutable data structures. Anyhow, I now only call eval once each time scorer is called, which is massively faster. This has allowed for additional testing points, and now I get a perfec

Re: regression

2009-05-26 Thread Daniel Lyons
On May 26, 2009, at 2:59 PM, kyle smith wrote: > > I have written code that will randomly guess the mathematical form of > a list of xy ordered pairs. My code and a sample run are in guess- > check.clj in the files section. The final sum of squares is > fantastic! This code is just for fun, b