Re: Mixed 64-bit system GerH binaries / BillYscripts --> two-sided training? YES!

I'm now wondering whether the idea of having more than 1 .CSS file is 
still a good idea...

I mean: as things stand now, we (crm114, that is) implicitly make a few 
assumptions when opening a certain .CSS file, like the class the tokens 
score should be attributed to (good/bad or perhaps 1/2/3/...) and a few 
other things.

Now, if we push forward the concept of two-sided training, why the heck 
shouldn't we merge the N .CSS files into a single, unified in 
fact, .CSS file?

What sounds clear enough to me is that the nasty 'the' word Ger 
complains about :-) would be present only once as a hash, and with a 
neutral score. Ah, the class the score is related to doesn't exist yet 
in the .CSS files, perhaps?

Hey, all we'd need to modify is "just" the whole .CSS concept, after 
all, and perhaps a few hundred thousand lines of (existing and working) 
code...  :-)

cheers,

Corrado


On Sunday 31 August 2008, Ger Hobbelt wrote:
[...]
> When you look at such CSS files and 'read' the files, you can say
> that in both CSS files the common word 'the' has an extremely high
> 'weight' (the 'count'!) which only 'equals out' if your
> classification takes the DIFFERENCE between left and right weight of
> a word[*] AND MOST IMPORTANTLY: you stuck religiously to the
> 'balanced training' dogma. [*]= (does _your_ classifier do that?
[...]
> The 'solution' (which will require extensive field testing, but I can
> hope, can I?):
>
> First of all, I need the training code to 'know about' the COMPLETE
> set of CSS files - as they are fed to the 'classify' statement as
> well. Right now, 'learn' only gets to see ONE of the buggers. Dang!
[...]

-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/