Raw data and wrapped class objects for the TransformedOutcome method are stored as class attributes. The wrapped class objects are described in the Usage: modeling section.
Everything else, from processed data to the transformation functions can be accessed as listed below:
up.randomized_search_params # Parameters that are used in `up.randomized_search()`
up.grid_search_params # Parameters that are used in `up.grid_search()`
up.transform # Outcome transform function.
up.untransform # Reverse of outcome transform function.
# Data (`y` in any of these can be replaced with `tc` for treatment or `x`).
up.transformed_y_train_pred # The predicted uplift.
up.transformed_y_train # The transformed outcome.
up.y_train
up.y_test
up.y # All the `y` data.
up.df
up.df_train
up.df_test
# Once a model has been created...
up.model
up.model_final
up.Q_cgains # 'aqini' or 'qini' can be used in place of 'cgains'
up.q1_cgains
up.q2_cgains
The raw data for all evaluation curves can be accessed within any UpliftEval object (upev below):
upev.PLOTTYPE_x # percentile
upev.PLOTTYPE_y
where the phrase PLOTTYPE can be replaced with any of the following: qini, aqini, cgains, cuplift, balance, uplift. Because up.test_results_ and up.train_results_ are UpliftEval class objects, they can also be similarly accessed as shown above.
The theoretical maximum curves can also be extracted:
# Overfitting theoretical maximal qini curve.
upev.qini_max_x # percentile
upev.qini_max_y
# "Practical" max curve.
upev.qini_pmax_x
upev.qini_pmax_y
# No sleeping dogs curve.
upev.qini_nosdmax_x
upev.qini_nosdmax_y
up.train_results_ can be used to plot the qini performance on the training data, as follows: up.train_results_.plot_qini().