Dummy Variables names as colum names in predicted dataset

moufkir
Level 2
Dummy Variables names as colum names in predicted dataset

Is their a way to get the exact column names the model used for dummification in the predicted data-set

dummy:variable_name:variable_value  or something like variable_name_variable_value ?

 

Thank you in advance!

 

2 Replies
DamienJ
Dataiker

Hi @moufkir,

You can't output those names in a dataset. However, there are available when you interpret some model results. For exemple: export coefficients of a logistic regression or export variable importance of a random forest.

You can also do the dummification yourself within a prepare recipe using the "unfold" preprocessor. Beware this might lead to huge datasets.

 

 

 

Damien Jacquemart, Lead Data Scientist @Dataiku
0 Kudos
moufkir
Level 2
Author

Thank you for your quick Answer.

for my case as the dummification is available in the model part, I think it is better to do it at this level.

and find it interesting to have an option to output dataset with dummified variables.

Best Regards  

0 Kudos