neural network - Rapidminer dummy coding mismatch -


i'm trying use neural network training on traindata , testing on testdata, do. however, data requires dummy coding of nominal features numerical. when that, trains neural network fails when applying test data (on apply exact same transformations/blocks) because of mismatch in dummy coding*.

*the error message in lines of: v47=h not exist in testdata

i checked , true testdata not have value 'h' @ in v47, while traindata has it. therefore, i'd ignore 'h' in v47, or replace it.

any way easily? keeping in mind might happen other features , going through features, 1 one, fix kind of issue, time consuming.

perhaps there's way tackle this?

thanks!

this similar previous post

this answer suggests combining test , training data cause possible values of nominal present splitting recover test , training sets again. possible additional nominal values retained in both splits.

this may not suit possibility use data weights operator on training example set. resulting weights can used select weights operator keep attributes of interest in test example set.


Comments

Popular posts from this blog

java - Run spring boot application error: Cannot instantiate interface org.springframework.context.ApplicationListener -

python - pip wont install .WHL files -

Excel VBA "Microsoft Windows Common Controls 6.0 (SP6)" Location Changes -