uplift

class skassist.models.uplift.Kane4(estimator)

Kane, et al., 2014

AB_Class==0: TN AB_Class==1: TR AB_Class==2: CN AB_Class==3: CR

class skassist.models.uplift.RealistUplift(estimator)

Modificatin of Shaar2016’s Reflective Uplift Model Assumption that all customers in CNR count towards the positive uplift is not realistc. Weight by overall uplift effect.

Supporting Features: ‘group’, ‘converted’

class skassist.models.uplift.ReflectiveUplift4(estimator)

Shaar2016, but with one multiclass model instead of two binary classifier. Supporting Features: ‘group’, ‘converted’

AB==0: TNR AB==1: TR AB==2: CNR AB==3: CR