Tabular - FillMissing for categorical

Part 1 (2019)
1

Just curious how the Tabular FillMissing proc handles categorical?
e.g.

df = pd.read_csv(path/'adult.csv')
df.loc[0:1000, 'workclass'] = np.nan

Would it just treat the nans as a class or use some other place holder e.g. unknown

2

It uses a third (if it’s T/F) category of unknown

1 Like
3

Thanks @muellerzr