This is the first time I'm participating in a ChaLearn AutoML challenge, so I might have just missed something in the documentation. If so, apologies.
When making a test submission, based on the information in the starting kit, I only had to upload the sample_code_submission folder.
I noticed that the data passed to Model in model.py already contains the data without missing values (imputation happens in the ingestion program by the datamanager).
I was wondering if as a participant you will never have access to the raw data, i.e. data will always be imputed for you?
Or is it possible to create a submission which handles data imputation itself?
Kind regardsPosted by: PGijsbers @ Aug. 23, 2018, 4:54 p.m.
Thanks for your question. Because of the format of AutoML3, participants will never have access to raw data, sorry
BestPosted by: hugo.jair @ Aug. 23, 2018, 10:12 p.m.