Re: VMs: Detecting "hands" automatically

[Bruce Grant <bgrant@xxxxxxxxxxxxx>]

Rene Zandbergen wrote:

> Can you tell me which file you used? This is
> very important.
>  
It is a reformatted version of the Currier ("C") transcription lines 
from the "intrln17.evt" interlinear file. I have been using it for a 
while so it may be a little out of date.

> Did you have only 145 pages, or did you select 
> only 145 pages?
>  
Only 145 of the pages were identified as A or B in my transcription file 
- there were 8 others with no hand (language) identification

> Does the algorithm tell you in the end whether the
> number (2) was a good choice or not?
No, it doesn't do that. You could probably get some idea by running it 
for different numbers of clusters and checking the numbers of members in 
the resulting clusters and/or the dispersion of  points from the cluster 
centers. Of course, the more clusters you allow the lower the dispersion 
will be, so I think it's a kind of a judgement call.

Regards,

Bruce


______________________________________________________________________
To unsubscribe, send mail to majordomo@xxxxxxxxxxx with a body saying:
unsubscribe vms-list