[mlpack] help with det

[Ryan Curtin]

On Sat, Feb 22, 2014 at 10:41:24AM +0000, Aakash Chaudhary wrote:
> can u help me with det (Density Estimation Trees) like from where to
> start.
>
> i have worked with few datasets and looked thorugh code.
> 
> just want to get through how it does and (what it does).

Hi Aakash,

Have you read through Pari's paper?

http://users.cis.fiu.edu/~lzhen001/activities/KDD2011Program/docs/p627.pdf

That should provide a better perspective on what density estimation
trees are for and how they differ from more conventional density
estimation techniques.

> when i tried it with small dataset (like iris) it was giving segmentation
> fault but its working with large datasets >5000 points (i think i m missing
> something about how testcases should be build for it)
> 
> should it be considered as a bug?

Yes, I think that is a bug, but can you provide more information on what
exactly you tried and what is going wrong?  That way we can start to
debug it.  Ideally, density estimation trees should work for even very
small datasets like the iris dataset, so if it is failing, filing a bug
on Trac (http://www.mlpack.org/trac/) is probably the right thing to do.

Thanks,

Ryan

-- 
Ryan Curtin    | "I am the luckiest man alive!"
ryan at ratml.org |   - General Borzov