I am also interested in Mondrian Forests (and partial_fit methods for things in general), though I thought one of the issues for implementing either of these methods was the way our trees are currently built would make it hard to extend to these two algorithms. It is definitely important not to regress performance of random forests after all the hard work that has gone on there.
I am not familiar enough with the cython code to really have a strong statement, though. On Thu, Mar 12, 2015 at 1:48 PM, Andreas Mueller <[email protected]> wrote: > Hi Meghana. > I'm -1 on adding decision jungles. > The main benefit seems to be faster predictions, something sklearn is not > really focusing on. > Also, it is a very recent publication, and I don't think it is widely used. > I think Mondrian Forests > http://arxiv.org/pdf/1406.2673v2 > https://github.com/balajiln/mondrianforest > would be much more interesting, but I also think they are not established > enough yet > (and the current implementation is python with MIT licence :) > > Cheers, > Andy > > > > On 03/12/2015 09:42 AM, meghana madhyastha wrote: > > Hi All, > > I am an undergraduate student who has worked on machine learning projects > using sklearn. I recently came across a paper published by a few researchers > Microsoft Research about decision jungles. To put it briefly, randomized > decision trees face a problem that given a lot of data, the number of nodes > in the decision trees will grow in depth. Decision Jungles addresses this > problem by introducing the idea of ensembles of DAGs. I thought that this > can be implemented in sklearn and I personally think its useful. I might be > applying for GSOC 15. > > -Meghana > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming The Go Parallel Website, > sponsored > by Intel and developed in partnership with Slashdot Media, is your hub for > all > things parallel software development, from weekly thought leadership blogs > to > news, videos, case studies, tutorials and more. Take a look and join the > conversation now. http://goparallel.sourceforge.net/ > > > > _______________________________________________ > Scikit-learn-general mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > > > > ------------------------------------------------------------------------------ > Dive into the World of Parallel Programming The Go Parallel Website, > sponsored > by Intel and developed in partnership with Slashdot Media, is your hub for > all > things parallel software development, from weekly thought leadership blogs > to > news, videos, case studies, tutorials and more. Take a look and join the > conversation now. http://goparallel.sourceforge.net/ > _______________________________________________ > Scikit-learn-general mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general > ------------------------------------------------------------------------------ Dive into the World of Parallel Programming The Go Parallel Website, sponsored by Intel and developed in partnership with Slashdot Media, is your hub for all things parallel software development, from weekly thought leadership blogs to news, videos, case studies, tutorials and more. Take a look and join the conversation now. http://goparallel.sourceforge.net/ _______________________________________________ Scikit-learn-general mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
