[RAS] test on ras.openlib.org

['Christian Zimmermann']

On Thu, 17 Nov 2011, Thomas Krichel wrote:

>  'Christian Zimmermann' writes
>
>> OK. But the learning on ACIS tries to separate Economics (say) from
>> not Economics.
>
>  Yes, but it can also separate various areas of economocs, or
>  different names of co-authors.
>
>> Here everything is Economics. Does that mean the
>> threshold should be set much stricter to put good matches in yes?
>
>  I don't know. It depends on the risk you want to take of false
>  positives and false negatives.

I guess we'll see from experience. Let's have rather strict for starters. 
Right now it is much too lax, as *everything* passes as 'yes'.

>
>
>  Cheers,
>
>  Thomas Krichel                    http://openlib.org/home/krichel
>                                      http://authorprofile.org/pkr1
>                                               skype: thomaskrichel
>

-- 
Christian Zimmermann                          FIGUGEGL!
Economic Research
Federal Reserve Bank of St. Louis
P.O. Box 442
St. Louis MO 63166-0442 USA
http://ideas.repec.org/zimm/