Re: VMs: Overfitting the Data (WAS: Another method different from Cardano Grilles)

[elvogt@xxxxxxxxxxx]

Zitat von Dennis <tsalagi@xxxxxxxx>:

> Jacques Guy wrote:
> > 
> 	Especially - how is overfitting the data different from 
> forming a hypothesis, aside from a hypothesis' ideally being 
> formulated before the fact?  

I guess it's all in the power of prediction.

You fit your data (design your wheels) based on half of your samples (one page 
of VM) until you're able to reproduce the data (the words). Then you check 
(without further fiddling your hypothesis), if you also manage to reproduce 
more samples (more pages of the VM).

In other words, if your disks/grilles/whatever which you designed according to 
analysis of one page of the VM are also able to reproduce the other 200 odd 
pages without further refinement, you're probably on to something...

Cheers,

   Elmar




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