Consider a dataset dat with only 3 binary variables (encoded using the LabelTransformer)
A
B
C
1
0
1
0
0
1
1
1
1
1
1
1
When I call synth.fit_sample(dat, transformer=tableTransformer) I seemingly get no errors for MST/AIM synthesizers. However, upon inspecting the input it looks a little something like this:
A
B
C
1
1
1
1
1
None
None
None
1
0
0
1
I don't suppose there is an easy fix for this, besides increasing sample size/epsilon. But at the very least I think fit_sample should generate a warning when None values are present.
Consider a dataset
dat
with only 3 binary variables (encoded using the LabelTransformer)When I call
synth.fit_sample(dat, transformer=tableTransformer)
I seemingly get no errors for MST/AIM synthesizers. However, upon inspecting the input it looks a little something like this:I don't suppose there is an easy fix for this, besides increasing sample size/epsilon. But at the very least I think
fit_sample
should generate a warning whenNone
values are present.