You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mahout.apache.org by Joe Kumar <jo...@gmail.com> on 2009/04/03 07:31:46 UTC
SVM implementation for Mahout
Hi all,
This is Joe. I am a part time computer science graduate student at UIUC.
Data Mining is one of my areas of research and interest.
I have applied for GSOC 2009 with a proposal to implement SVM.
I saw a Jira issue (Mahout-14) which had a conversation on trying to reuse
svmlin. I was wondering if anyone had already contacted Vikas (author of
svmlin). If not, I can check with him to see if we could reuse svmlin.
I can otherwise work on implementing SVM from scratch. Although it might be
like reinventing the wheel, I guess it is better to implement core
algorithms like SVM by ourselves.
would appreciate your thoughts and guidance.
thanks
Joe.
Re: SVM implementation for Mahout
Posted by Joe Kumar <jo...@gmail.com>.
Ted,
Thanks for your feedback and thoughts. those f/b make sense.
Isabel,
I havent specifically probed Vikas about joining Mahout but he is very much
willing to help us out in any way he can. I'll probably check if he would be
interested in joining Mahout. It sure would be good to have him on the team.
On Tue, May 5, 2009 at 1:12 AM, Isabel Drost <is...@apache.org> wrote:
> On Monday 04 May 2009 06:39:45 Ted Dunning wrote:
> > In the meantime, whatever is the easiest way to make forward progress is
> > probably your best bet.
>
> +1 from me.
>
> Joe did you talk to Vikas about joining the Mahout project? I would prefer
> integrating not only the code but also getting the developers in - provided
> they have similar goals to what Mahout wants to achieve.
>
> Isabel
>
>
>
Re: SVM implementation for Mahout
Posted by Isabel Drost <is...@apache.org>.
On Monday 04 May 2009 06:39:45 Ted Dunning wrote:
> In the meantime, whatever is the easiest way to make forward progress is
> probably your best bet.
+1 from me.
Joe did you talk to Vikas about joining the Mahout project? I would prefer
integrating not only the code but also getting the developers in - provided
they have similar goals to what Mahout wants to achieve.
Isabel
Re: SVM implementation for Mahout
Posted by Ted Dunning <te...@gmail.com>.
I think that the primary criterion is that we want to provide capabilities
for users of Mahout. Providing an SVM implementation would be the first
priority, regardless of source (presuming licensing and such is handled
correctly).
That said, the rub is likely to come when a scalable implementation is to be
produced. SVM isn't nearly as simple to parallelize as some other
algorithms so it may well become necessary to make some pretty substantial
modifications.
In the meantime, whatever is the easiest way to make forward progress is
probably your best bet.
On Sun, May 3, 2009 at 7:45 PM, Joe Kumar <jo...@gmail.com> wrote:
> I just got back from Vikas reg svmlin and he is ok to use apache 2.0
> license.
>
> That aside, I thought of sharing some of my perspectives.
> SVM is one of the widely used classification algorithms and I guess it is
> better to have our own implementation. We could reuse existing libraries
> for
> non-core algorithms or components. Assuming that SVM on Mahout would be
> quite widely used, would it be better to write our own ?
>
> On Mon, Apr 6, 2009 at 1:58 AM, Sean Owen <sr...@gmail.com> wrote:
>
> > Apache 2.0 License would be easiest I imagine?
> >
> > On Mon, Apr 6, 2009 at 6:12 AM, Joe Kumar <jo...@gmail.com> wrote:
> > > He was asking about what licensing he should specify for svmlin so that
> > it
> > > can used in mahout ? any thoughts / suggestions please
> >
>
--
Ted Dunning, CTO
DeepDyve
111 West Evelyn Ave. Ste. 202
Sunnyvale, CA 94086
www.deepdyve.com
858-414-0013 (m)
408-773-0220 (fax)
Re: SVM implementation for Mahout
Posted by Joe Kumar <jo...@gmail.com>.
I just got back from Vikas reg svmlin and he is ok to use apache 2.0
license.
That aside, I thought of sharing some of my perspectives.
SVM is one of the widely used classification algorithms and I guess it is
better to have our own implementation. We could reuse existing libraries for
non-core algorithms or components. Assuming that SVM on Mahout would be
quite widely used, would it be better to write our own ?
On Mon, Apr 6, 2009 at 1:58 AM, Sean Owen <sr...@gmail.com> wrote:
> Apache 2.0 License would be easiest I imagine?
>
> On Mon, Apr 6, 2009 at 6:12 AM, Joe Kumar <jo...@gmail.com> wrote:
> > He was asking about what licensing he should specify for svmlin so that
> it
> > can used in mahout ? any thoughts / suggestions please
>
Re: SVM implementation for Mahout
Posted by Sean Owen <sr...@gmail.com>.
Apache 2.0 License would be easiest I imagine?
On Mon, Apr 6, 2009 at 6:12 AM, Joe Kumar <jo...@gmail.com> wrote:
> He was asking about what licensing he should specify for svmlin so that it
> can used in mahout ? any thoughts / suggestions please
Re: SVM implementation for Mahout
Posted by Joe Kumar <jo...@gmail.com>.
Thanks Isabel. I communicated with Vikas and he is glad to relax his license
so as to reuse svmlin in mahout.
He was asking about what licensing he should specify for svmlin so that it
can used in mahout ? any thoughts / suggestions please
I am new to Hadoop and I was wondering if svmlin would be a good choice for
exploiting the power of Hadoop ? has anyone already analyzed on this ? would
appreciate any pointers
thanks
Joe.
On Sun, Apr 5, 2009 at 7:05 AM, Isabel Drost <is...@apache.org> wrote:
> On Friday 03 April 2009, Joe Kumar wrote:
> > I was wondering if anyone had already contacted Vikas (author of
> > svmlin). If not, I can check with him to see if we could reuse svmlin.
>
> I have contacted Vikas quite a while ago. He even seemed interested in
> Mahout
> and offered to contribute to the project. Unfortunately I have not heard
> back
> from him for several months and have not seen him on our lists :( Feel free
> to ping him concerning svmlin.
>
> Isabel
>
> --
> Democracy is good. I say this because other systems are worse. --
> Jawaharlal
> Nehru
> |\ _,,,---,,_ Web: <http://www.isabel-drost.de>
> /,`.-'`' -. ;-;;,_
> |,4- ) )-,_..;\ ( `'-'
> '---''(_/--' `-'\_) (fL) IM: <xm...@spaceboyz.net>
>
--
God doesnot ask us of our ability but our availability
http://joekumar.net
Re: SVM implementation for Mahout
Posted by Isabel Drost <is...@apache.org>.
On Friday 03 April 2009, Joe Kumar wrote:
> I was wondering if anyone had already contacted Vikas (author of
> svmlin). If not, I can check with him to see if we could reuse svmlin.
I have contacted Vikas quite a while ago. He even seemed interested in Mahout
and offered to contribute to the project. Unfortunately I have not heard back
from him for several months and have not seen him on our lists :( Feel free
to ping him concerning svmlin.
Isabel
--
Democracy is good. I say this because other systems are worse. -- Jawaharlal
Nehru
|\ _,,,---,,_ Web: <http://www.isabel-drost.de>
/,`.-'`' -. ;-;;,_
|,4- ) )-,_..;\ ( `'-'
'---''(_/--' `-'\_) (fL) IM: <xm...@spaceboyz.net>