#playground-series-s3e26

1 messages · Page 1 of 1 (latest)

stoic trench
stoic trench
barren scaffold
#

i can not seem to increase my competition score and i have tried hyperparameter tuning, cross validation, scaling data, feature engineering and early stopping
what else should try?

stoic trench
#

Facing same issue

fervent pike
barren scaffold
fervent pike
barren scaffold
#

yea

fervent pike
royal mortar
#

Hi, I'm currently going crazy, I have these two codes that should do the same thing :

#

If I call it like that (using the preprocessor) I get a bad score

#

while if I use the preprocess function on the data, I get a much better score

#

These code seem to be doing EXACTLY the same thing, I am really confused

#

Does anyone have an idea of what is happening ? 😦

rocky ruin
#

@royal mortar what do your resultant data frames look like?

royal mortar
#

The one resulting from the preprocess function

#

Not sure how to show you the preprocessor one as calling fit_transform doesn't return a data frame :/

#

well I can convert it but I'll lose all column names

#

Ok found the mistake :

When using ColumnTransformer, the columns that are not specified in the set of transformers are removed

#

I need to add a transformer that does... nothing

royal mortar
#

thanks!

wary grove
#

why do ppl delete their kaggle accounts

stoic trench
#

I dont know