Draws a waterfall chart cascading from baseline_value to the final prediction (hc_information_visualization.md SS5: "correct for local explanation with
directional attribution... bars cascade from base rate to prediction; color
encodes direction; length encodes magnitude"). This is the one bar chart in this module that legitimately does not start at zero – Tufte's lie-factor rule is satisfied relative to baseline_value, the meaningful reference point, not zero.
More...
#include <attribution_waterfall.hpp>
|
| static void | Draw (const char *title, const Attribution &attr, float baseline_value) |
| |
Draws a waterfall chart cascading from baseline_value to the final prediction (hc_information_visualization.md SS5: "correct for local explanation with
directional attribution... bars cascade from base rate to prediction; color
encodes direction; length encodes magnitude"). This is the one bar chart in this module that legitimately does not start at zero – Tufte's lie-factor rule is satisfied relative to baseline_value, the meaningful reference point, not zero.
- Note
- Deliberately not unit-tested – see AttributionBarChart's note. All real logic (cumulative computation) lives in plot_data.hpp's ToWaterfallSteps, which is.
◆ Draw()
| static void pulsatrix::AttributionWaterfallChart::Draw |
( |
const char * |
title, |
|
|
const Attribution & |
attr, |
|
|
float |
baseline_value |
|
) |
| |
|
static |
- Parameters
-
| title | ImPlot plot title/id. |
| attr | A rank-1 Attribution, in the explainer's own feature order. |
| baseline_value | The starting reference value (e.g. IG's baseline prediction). |
The documentation for this class was generated from the following file: