Issue with type of scaling when calling get_predictor().preprocess(df)

sourjeet
Level 1
Issue with type of scaling when calling get_predictor().preprocess(df)

Hi All,

I have trained a model with scaling as min-max scalling for all features.

When I am reading the same model and feeding the train data (df) using the following code : 

model_minMax = dataiku.Model(<model name>) # Provide model name
my_predictor = model_minMax.get_predictor()

xi = my_predictor.preprocess(df)[0]

pd.DataFrame(xi).describe()

 

I see that the scaling is done as average-scaling. 

Request you to kindly help.

 

Attaching log of preprocessing.

 

Regards

Sourjeet

0 Kudos
2 Replies
AllanF
Dataiker

Hello Sourjeet,

Thank you for your message, you actually found a bug in DSS ๐Ÿ™‚

We logged it and it will be fixed in a future release.

Samuel_R_
Dataiker

Hi @sourjeet, the forthcoming Dataiku 9.0.2 release will fix this bug. Thanks for reporting it !