[SciPy-user] Bayes net question

[David Warde-Farley]

On 16-Jan-08, at 12:05 PM, Karl Young wrote:

> I've looked at the documentation for that package (but refuse to go  
> back
> to Matlab re. actually using it :-)) and it looks like it's quite a  
> bit
> further along than OpenBayes re. options so does it seem like a
> consensus that we should wait for this to be settled before any
> decisions about incorporating Bayes nets ? As I mentioned before I'm
> willing to help (e.g. with a port) once a decision is reached re. the
> best approach.

Likewise, I'd be willing to help with a port.

I know that BNT includes code from other projects (for one thing the  
IRLS implementation for optimizing the weights of a softmax from the  
NetLab neural networks toolbox), so it may not be totally up to  
Professor Murphy whether or not to dual-license the entire thing.

DWF