That having been said, however - it is probably a good idea for us to write an adapter which allows the random forest code to read distributed sparse matrix formats, as that is a shared format plenty of other parts of the project can use...
I smell a JIRA ticket coming on (feel free to file it, Karan, and I will make sure it doesn't get completely overlooked!)... -jake On Mon, Jun 7, 2010 at 11:13 PM, deneche abdelhakim <[email protected]>wrote: > Hi Karan, > > For now, the only way is to write your data in a text file (in a format > similar to UCI Datasets) and then pass this file to the random forest. > > --- En date de : Lun 7.6.10, Karan Jindal < > [email protected]> a écrit : > > > De: Karan Jindal <[email protected]> > > Objet: Re: Reg: Building Decision Forest using Mahout 0.3 > > À: [email protected] > > Date: Lundi 7 juin 2010, 6h03 > > Hi deneche abdelhakim, > > > > If I want to implement a new input format for the random > > forest say which > > take sparse matrix as a input, How should I proceed in that > > direction? > > > > Any ideas?? > > > > Regards > > Karan > > > > > > > > > yes thats the documentation for Mahout's Decision > > Forests, but as Ted > > > noted before, you'll need to checkout the latest > > version of the trunk. > > > > > > Please note that it's still a work in progress, I > > would appreciate your > > > feedback and suggestions. Feel free to ask questions > > in this mailing list > > > > > > --- En date de : Lun 31.5.10, Karan Jindal > > > <[email protected]> > > a écrit : > > > > > >> De: Karan Jindal <[email protected]> > > >> Objet: Re: Reg: Building Decision Forest using > > Mahout 0.3 > > >> À: [email protected] > > >> Date: Lundi 31 mai 2010, 13h51 > > >> Hi Ted, > > >> > > >> Thanks for the advice of using the latest version > > of > > >> mahout. > > >> It worked for me :) . > > >> > > >> And also I followed this link > > >> > https://cwiki.apache.org/confluence/display/MAHOUT/Partial+Implementation > > >> for running the mahout. > > >> > > >> > > >> > > >> > Mahout supports the construction of Random > > Forests > > >> which are a form of > > >> > decision forest. > > >> > > > >> > It would be much easier for you to get > > started if you > > >> could upgrade to the > > >> > current latest development version. Version > > 0.3 > > >> is the latest stable > > >> > version, but lots of stuff has changed for > > the better > > >> since 0.3. > > >> > > > >> > In any case, the class that you are looking > > for is > > >> > org.apache.mahout.df.mapred.BuildForest. > > >> > > > >> > On Sun, May 30, 2010 at 10:08 PM, Karan > > Jindal < > > >> > [email protected]> > > >> wrote: > > >> > > > >> >> > > >> >> Hi all, > > >> >> > > >> >> I had installed the latest version of > > mahout i.e. > > >> 0.3. > > >> >> I want to build a decision forest, but > > didn't find > > >> any documentation > > >> >> which > > >> >> guides to me to so. > > >> >> > > >> >> Can anyone tell me where to find such > > documentaion > > >> which will tell about > > >> >> how to run mahout on hadoop and also > > various > > >> command line options for > > >> >> Building Decision Tree? > > >> >> > > >> >> Regards > > >> >> Karan > > >> >> > > >> > > > >> > > >> > > > > > > > > > > > > > > > -- > > > This message has been scanned for viruses and > > > dangerous content by MailScanner, and is > > > believed to be clean. > > > > > > > > > > > > -- > > This message has been scanned for viruses and > > dangerous content by MailScanner, and is > > believed to be clean. > > > > > > > >
