search

ltflores / csc-869-mlog

Automatically exported from code.google.com/p/csc-869-mlog
0 stars 0 forks source link

Exclude attributes to trace author from classification #2

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Exclude attributes added by the SentenceBasedTextDirectoryLoader.java like 
democrats_03152011/JamesMcDermott or republicans_03152011/MarioDiaz-Balart from 
the classification. Optionally just disable the generation of them as they're 
not used at the moment anyways

Original issue reported on code.google.com by markus.neubrand on 5 Apr 2011 at 11:53

GoogleCodeExporter commented 9 years ago 
Just realized we need this information for the custom cross-validation. Don't 
remove those author attributes!

Original comment by markus.neubrand on 5 Apr 2011 at 11:56

GoogleCodeExporter commented 9 years ago

Original comment by markus.neubrand on 6 Apr 2011 at 12:09

GoogleCodeExporter commented 9 years ago 
Changed setOutputFilename(true) to setOutputFilename(false), the original 
comment in our code said it was set to true to "save filename" but it really 
adds the filename as an attribute.

Original comment by ltflo...@gmail.com on 7 Apr 2011 at 4:36

GoogleCodeExporter commented 9 years ago

Original comment by ltflo...@gmail.com on 7 Apr 2011 at 4:36

GoogleCodeExporter commented 9 years ago 
K, now I see what we were trying to do with setOutputFilename(true), for now we 
should leave it as setOutputFilename(false), but if we want to ensure that we 
cross validate by senator and not just tweets we'll need to find a way to keep 
this info with instances.

Original comment by ltflo...@gmail.com on 7 Apr 2011 at 6:15