Run Best Tabular Classifier on Same Observations as Graph NN

[ijyliu]

If time allows, try to make a performance comparison.

Rerunning would only be involved if we wanted to limit the training data of the tabular classifier to just calls in the training graph - otherwise, we can just filter the predictions to those that we have predictions from the graph NN for.

[ijyliu]

do for logistic regression too, see if in between lr and xgboost

[ijyliu]

completed

[ijyliu]

need to run mcnemars comparisons

[ijyliu]

done