H2O have a very high quality random high performance implementation of
random forests.
On Mon, Aug 11, 2014 at 1:49 PM, Sameer Tilak wrote:
>
>
>
>
>
>
>
>
> Hi All,
> We are currently using Weka. I looked the the site and read briefly about
> experimental nature of Mahout on Spark. I was wond
my problem.
>
> > Date: Mon, 11 Aug 2014 14:01:12 -0700
> > Subject: Re: Mahout on Spark: random forest
> > From: dlie...@gmail.com
> > To: user@mahout.apache.org
> >
> > I am not sure they have the forests yet. They have what seems to be a
> > decently perfor
Yes, I have started working with that since morning and see how well it works
for my problem.
> Date: Mon, 11 Aug 2014 14:01:12 -0700
> Subject: Re: Mahout on Spark: random forest
> From: dlie...@gmail.com
> To: user@mahout.apache.org
>
> I am not sure they have the fore
I am not sure they have the forests yet. They have what seems to be a
decently performant decision tree tho.
On Mon, Aug 11, 2014 at 1:54 PM, Suneel Marthi
wrote:
> there is no Random Forest impl on Spark in Mahout yet. Ml-lib has a Random
> Forests impl why can't u use that instead.
>
>
> On
>From what I read on the mailing lists was that the rf implementation is on the
>roadmap and only decision tree support is included at present. However, I may
>be completely wrong.
> Date: Tue, 12 Aug 2014 02:24:21 +0530
> Subject: Re: Mahout on Spark: random forest
>
there is no Random Forest impl on Spark in Mahout yet. Ml-lib has a Random
Forests impl why can't u use that instead.
On Tue, Aug 12, 2014 at 2:19 AM, Sameer Tilak wrote:
>
>
>
>
>
>
>
>
> Hi All,
> We are currently using Weka. I looked the the site and read briefly about
> experimental natur
afaik there's no RF in Mahout on Spark. But there's something that may be
pretty close, in Spark's mlib.
On Mon, Aug 11, 2014 at 1:49 PM, Sameer Tilak wrote:
>
>
>
>
>
>
>
>
> Hi All,
> We are currently using Weka. I looked the the site and read briefly about
> experimental nature of Mahout o
Hi All,
We are currently using Weka. I looked the the site and read briefly about
experimental nature of Mahout on Spark. I was wondering how mu of Mahout's
functionality is available currently? For example, I am interested in using
random forest implementation. I wanted to check if I c
