healthcareai-r icon indicating copy to clipboard operation
healthcareai-r copied to clipboard

prep_data does not remove the ignored variables

Open AviralVijay-GSLab opened this issue 6 years ago • 0 comments

machine_learn function uses the prep_data function to prepare the data which is feed to the tune_model or flash model later. prep_data function have !!!dots or ignored variable which should not be included in model training, but prep_data gives the ignored columns also in the resultant dataset, on the other hand if one wants to use the same recipe then it removes the ignored columns.

Feature Requests

machine_learn function have impact of this issue because it directly provide the pd(i.e outcome of prep_data) to the tune_models or flash_models function, where ignored columns also be used in model training, that is controversial with the description provided in the model parameter and also with the applications of this function.

@michaellevy Please find the attached below screenshots and pdf document for more information.

Feature_Request.pdf

machine_learn 1

machine_learn 2

AviralVijay-GSLab avatar Jul 30 '18 09:07 AviralVijay-GSLab