Mailing-List: contact user-help@mahout.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@mahout.apache.org
Received-SPF: pass (athena.apache.org: domain of ted.dunning@gmail.com
 designates 209.85.216.170 as permitted sender)
MIME-Version: 1.0
In-Reply-To: <3be90d3b.73ed.1321114aa7f.Coremail.myn@163.com>
References: <3be90d3b.73ed.1321114aa7f.Coremail.myn@163.com>
Date: Sun, 28 Aug 2011 09:02:32 -0700
Message-ID: 
 <CAJwFCa1MQKEGRA3F2Vqcj7uZ0VcKcQv6JYnzV__EfO8J64U1hw@mail.gmail.com>
Subject: Re: why limit the num of numCategories in
 AbstractOnlineLogisticRegression
From: Ted Dunning <ted.dunning@gmail.com>
To: "user@mahout.apache.org" <user@mahout.apache.org>
Content-Type: multipart/alternative; boundary=20cf303a2c0ff5c10904ab92e779

--20cf303a2c0ff5c10904ab92e779
Content-Type: text/plain; charset=UTF-8

That limit is only for that one classifier call. The idea is that if you
have two categories then getting just one score eliminates the construction
of a vector. Use classify() In the general case.

On Sunday, August 28, 2011, myn <myn@163.com> wrote:
> I was read the AbstractOnlineLogisticRegression(SGD) sorce code  ,but
> it limit the number of numCategories to 2,why? why not 3 or more?
>
> @Override
>  public double classifyScalar(Vector instance) {
>    Preconditions.checkArgument(numCategories() == 2, "Can only call
classifyScalar with two categories");
>
>    // apply pending regularization to whichever coefficients matter
>    regularize(instance);
>
>    // result is a vector with one element so we can just use dot product
>    return link(classifyScalarNoLink(instance));
>  }
>

--20cf303a2c0ff5c10904ab92e779--