info_plots.actual_plot() got an error

[fenxouxiaoquan]

when i execute the follow code just like binary_classification tutorial:
"""
fig, axes, summary_df = info_plots.actual_plot(
model=titanic_model, X=titanic_data[titanic_features], feature=['Embarked_C', 'Embarked_S', 'Embarked_Q'],
feature_name='embarked'
)
"""
and got follow error:
"""
TypeError: predict_proba() argument after ** must be a mapping, not NoneType
"""
i also tried lgb.LGBMClassifier and lgb raw model on my own data but got same error.
is there anyone knows how to fix it?

[vickiwickinger]

Hi,

the error message points to the following line in the source code
prediction = predict(X, **predict_kwds)
I just changed this line to prediction = predict(X) and then it worked again. Probably the scikit API changed in recent updated and it's not possible anymore to overload an additional None value.

[mfgray]

I ran in to this issue too. The info_plots.actual_plot() has the keyword argument predict_kwds which has the default value None, which seems to be causing the error with the argument unpacking (**). I was getting the same error.
"""
predict() argument after ** must be a mapping, not NoneType
"""

I was able to get it to work by passing in an empty set as the keyword argument. I.e. when calling actual_plot pass in the following keyword arg:
predict_kwds={}

Looking at the git commits it looks like the author had recently changed the default from an empty set to None.