mahout-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Dunning <ted.dunn...@gmail.com>
Subject Re: [jira] [Commented] (MAHOUT-479) Streamline classification/ clustering data structures
Date Thu, 16 Jun 2011 22:36:07 GMT
THis is correct behavior (except for the typo where you say <<0 ... a radius
should always be > 0)

The reason that this works out is that the prior should set the probability
of a tiny cluster to be near zero.

On Thu, Jun 16, 2011 at 11:05 AM, Vasil Vasilev (JIRA) <jira@apache.org>wrote:

> 2. dNorm returns probability density, not probability, which means that for
> the cases where radius << 0 and the number of dimensions of the feature
> vectors is very big (~50000) the pdf goes quickly to infinity.
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message