[SciPy-user] NNLS in scipy
Sun Jun 15 16:45:08 CDT 2008
I was just going to do a blog post on this same topic next week. I have a
few different versions of python NMF code I used a few years ago, including
sparse, parallel, and memmap versions.
Until I post my examples, you can find a basic implementation based on scipy
here (from Toby Segaran):
These guys also did an implementation in python:
Some more NMF related links here:
On Sun, Jun 15, 2008 at 12:07 PM, Christian Meesters <email@example.com>
> I'd like to perform some "non-negatively constraint least squares"
> algorithm to fit my data, like this:
> Signal = SUM a_i C_i
> where C_i is some simulated signal and a_i the amplitude contributed by
> that simulated signal. Or in terms of arrays, I will have one reference
> array and several arrays of simulated signals. How can find the
> (non-negative) coefficients a_i for each simulated signal array? (All
> negative contributions should be discarded.)
> Is there anything like that in scipy (which I couldn't find)? Or any
> other code doing that?
> Else I could write it myself and contribute, but having some working
> code would be nice, of course.
> SciPy-user mailing list
Peter N. Skomoroch
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the SciPy-user