in

.NET Opensource

Community for opensource projects by Christoph Rüegg

Non-negative linear regression and F-distribution?

Last post 03-31-2008 12:06 by Christoph Rüegg. 4 replies.
Page 1 of 1 (5 items)
Sort Posts: Previous Next
  • 02-03-2008 17:11

    Non-negative linear regression and F-distribution?

     I can use Matrix.Solve to do linear regression. But, is it possible to do non-negative linear regression? Actually, I have a code converted from c++. But if there is such one, I'd love to use it directly.

     And, after linear regression, I need to do F-test. Is there a F-distribution? Thanks.

  • 03-12-2008 16:51 In reply to

    Re: Non-negative linear regression and F-distribution?

     For the second part: The F-Distribution (Fisher-Snedecor) was just implemented in revision 383, see IRID-108.

  • 03-18-2008 17:05 In reply to

    Re: Non-negative linear regression and F-distribution?

    Christoph: first thanks for your excellent work.

     Is it possible to actually download the latest revision? I need to do calculate some F-probabilities; tried it using the beta function (the formula found on wikipedia) but it does not yield the correct results. Any insights? Thanks!

  • 03-20-2008 11:17 In reply to

    Re: Non-negative linear regression and F-distribution?

    I plan to release the upcoming Iridium iteration 12 before end of March. Let me know if you need it faster, I can of course email you an intermediate build (or you can try to build it yourself).

  • 03-31-2008 12:06 In reply to

    Re: Non-negative linear regression and F-distribution?

    The Iridium 2008 March Release (v2008.3.12.405) is now available, which includes the F-Distribution. Let me know if you observe the same issues with our implementation. The probability density function is unit tested, but we can only test a limited set of parameters...
Page 1 of 1 (5 items)
Powered by Community Server (Non-Commercial Edition), by Telligent Systems