> 
> Thanks for the fast response.
> 
> 
> to JP: It works for me using gcc and g++ on 32-bit Mac and Linux! :)
> 
> 
> J. Friedman in the paper "Greedy Function Approximation: A Gradient
> Boosting Machine" has mentioned the M-regression algorithm which is
> a gradient boosting regression method with huber loss function.
> 

Does it work well? Is it widely used? If so, pull requests are always welcome - 
as long as they fit nicely in the current interface.

> 
> Adding these features is generally useful for the users and makes the
> library a unified framework for machine learning. I need structured
> output svm for plane parameter estimation from 3d depth points and
> gradient boosting for a similar task but want to minimize my own
> loss function.

Have you thought about how to formulate your task as a SSVM?

Cheers,
Andy

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to