[Corpora-List] Metrics for corpus "parseability"

[Max Chevalier]

Dear All,

I am a new user of this list....

I wonder if someone know some techniques to evaluate the content 
homogeneity of a corpora. That is to say that I would evaluate the 
number (few or a lot) of themes in documents....

Is anyone has some idea?

Sincerely yours,

Max.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: chevalie.vcf
Type: text/x-vcard
Size: 825 bytes
Desc: not available
URL: <http://listserv.linguistlist.org/pipermail/corpora/attachments/20080204/3c6ea0d8/attachment-0001.vcf>
-------------- next part --------------
_______________________________________________
Corpora mailing list
Corpora at uib.no
http://mailman.uib.no/listinfo/corpora