fix_missing possible bug? about fastai

[jph00]

Yes it could do, I think. I guess give it a go and see if you can come up with a situation where it fails - if you can, a PR to fix would be appreciated! :)

[AdityaSoni19031997]

We might need to look for nan's in test set also?

[EKami]

We might need to look for nan's in test set also?

Yes, NaN could exists in the test set and not in the train set and given the actual code that would cause some issues regarding the resulting dimensions of the 2 sets.

[AdityaSoni19031997]

Actually I was always under the impression that the test set will atleast won't contain that or will have a method to auto resolve that issues(using imputer like thing) or remove that particular entry..
Thanks.

[jph00]

I'll close this in the absence of a failing test case - feel free to reopen if you manage to create one.

[AdityaSoni19031997]

Can we use proc_df in the testset?
It seems that y_fld is to be mentioned?

[AdityaSoni19031997]

Ran into this issue while trying out the kaggle's House Regression Price Prediction Competition