+1. But I would rather try to convince the Mahout guys to implement Machine learning algorithms in BSP for the long run. Just to teaser people, we can try to make the k-means a bit more user friendly and use it as an example.
For the most machine learning problems we could need a matrix operation package as well, this is on our roadmap for months and years. Just basic multiplication and transposing should be enough. 2011/12/22 Edward J. Yoon <[email protected]> > +1 > > On Thu, Dec 22, 2011 at 5:36 PM, Tommaso Teofili > <[email protected]> wrote: > > Hi all, > > following recent work from Thomas about K-Means algorithm for clustering > > I'd like to propose a dedicated package for BSP based machine learning > > algorithms implementations under hama-examples module. > > WDYT? > > All the best. > > Tommaso > > > > -- > Best Regards, Edward J. Yoon > @eddieyoon > -- Thomas Jungblut Berlin <[email protected]>
