Naming conventions
rtichoke uses a consistent naming pattern to distinguish high-level
workflows from lower-level plotting and rendering functions. Functions
named create_*() start from predictions and outcomes and
create a curve, table, or report.
prepare_performance_data() exposes the common
performance-data layer, while plot_*() and
render_*() functions operate on already prepared
performance data where those lower-level interfaces are available.
The table below summarizes the main entry points.
| Predictions and Outcomes | Performance Data | |
|---|---|---|
| Performance Data | prepare_performance_data() |
|
| ROC | create_roc_curve() |
plot_roc_curve() |
| Lift | create_lift_curve() |
plot_lift_curve() |
| Gains | create_gains_curve() |
plot_gains_curve() |
| Precision Recall | create_precision_recall_curve() |
plot_precision_recall_curve() |
| Decision | create_decision_curve() |
plot_decision_curve() |
| Calibration | create_calibration_curve() |
|
| Performance Table | create_performance_table() |
render_performance_table() |
| Summary Report | create_summary_report() |