Dear Luca and others,
Logarithm of probability bit-costing can be used for cluster
validation, decision tree/graph error rates and all sorts of validations
and measurements of error.
For more information, see section 6 of
Y. Agusta and D. L. Dowe (2003b). Unsupervised Learning of Correlated
Multivariate Gaussian Mixture Models using MML, (to appear) In Proc.
16th Australian Joint Conference on Artificial Intelligence (AI'03),
Perth, Australia, 3-5 Dec. 2003
(downloadable from
http://www.csse.monash.edu.au/~dld/Publications/2003/Agusta+Dowe2003b.ref )
and possibly also section 5.1 and table 3 of
P. J. Tan and D. L. Dowe (2003). MML Inference of Decision Graphs
with Multi-Way Joins and Dynamic Attributes, (to appear) In Proc.
16th Australian Joint Conference on Artificial Intelligence (AI'03),
Perth, Australia, 3-5 Dec. 2003
(downloadable from
http://www.csse.monash.edu.au/~dld/Publications/2003/Tan+Dowe2003.ref ).
David.
> From [EMAIL PROTECTED] Tue Sep 23 16:25:57 2003
> Subject: Re: cluster validation criteria
> To: [EMAIL PROTECTED]
>
> This is a multipart message in MIME format.
>
> --Boundary_(ID_EVUiIGySfGHKYw39GktAWQ)
> Content-type: text/plain; charset=us-ascii
> Content-transfer-encoding: 7BIT
>
> There are several measures for cluster validity.
> The following letter to the editor could be a starting point for you
> http://www3.interscience.wiley.com/cgi-bin/abstract/78003403/ABSTRACT
> Hope it helps
> Regards
>
> Luca
>
> --Boundary_(ID_EVUiIGySfGHKYw39GktAWQ)
