Mailing-List: contact user-help@mahout.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@mahout.apache.org
Received-SPF: pass (nike.apache.org: domain of oeddyo@gmail.com designates
 209.85.160.42 as permitted sender)
DomainKey-Signature: a=rsa-sha1; c=nofws;
        d=gmail.com; s=gamma;
        h=mime-version:in-reply-to:references:from:date:message-id:subject:to
         :cc:content-type;
        b=K4IZtdmOUbeqF8R769qwuHL+YNyUExHEssqaztXm8/yOcVkzZYtsFj9Cur+T9vW5yf
         lytWkQu41WLLRf9zv39pyILGvQeYDbeYDPunhscmsRRgRcGK33nNcGw5DL4vlJjMYZGM
         y5JdiiAlv7eMgVW2vfJirP+84kDhLGLxHFejg=
MIME-Version: 1.0
In-Reply-To: <BANLkTinx8yKW=eJ8sFdP4KKD86TfzrNfXA@mail.gmail.com>
References: <BANLkTikjLSvN_BZSUUbGOG-DbBdUwg9J=A@mail.gmail.com>
 <BANLkTinx8yKW=eJ8sFdP4KKD86TfzrNfXA@mail.gmail.com>
From: ke xie <oeddyo@gmail.com>
Date: Wed, 13 Apr 2011 15:43:58 +0800
Message-ID: <BANLkTi=Sc0x3gDRa77QPdsmSUiZvE0-zSg@mail.gmail.com>
Subject: Re: How about a LSH recommender ?
To: Ted Dunning <ted.dunning@gmail.com>
Cc: user@mahout.apache.org
Content-Type: multipart/alternative; boundary=000e0cd1876ae155e304a0c7f961

--000e0cd1876ae155e304a0c7f961
Content-Type: text/plain; charset=ISO-8859-1

Ok, I would try to implement a none-distributed one. Actually I have a
python version now.

But I have a problem. When doing min-hash, the matrix should be either 1 or
0, and then do the hash functions. Then how about rating data? If the matrix
is filled with 1~5 numbers, should we convert them use some treshould and
convert the rating to 1 if the rating is more than the treshould?

This is the reference I read about LSH. check it out (chapter 3)
http://infolab.stanford.edu/~ullman/mmds.html

On Wed, Apr 13, 2011 at 3:25 PM, Ted Dunning <ted.dunning@gmail.com> wrote:

> Sure.
>
> LSH is a fine candidate for parallelism and scaling.
>
> I would recommend starting small and testing as you go rather than leaping
> into a parallelized full-fledged implementation.  Look for other open-source
> implementaions of LSH algorithms.
>
> Be warned that the parameter selection for LSH can be pretty tricky (so I
> hear, anyway).  You should pick a reasonable and realistic test problem so
> that you can experiment with that.
>
>
> On Wed, Apr 13, 2011 at 12:19 AM, ke xie <oeddyo@gmail.com> wrote:
>
>> Can we implement one and contribute into the mahout project? Any
>> suggestions?
>>
>
>


-- 
Name: Ke Xie   Eddy
Research Group of Information Retrieval
State Key Laboratory of Intelligent Technology and Systems
Tsinghua University

--000e0cd1876ae155e304a0c7f961--