Getting started#

For direct arrays, the KDS functions are the shortest path:

from scikitplot.decile.kds import decile_table

table = decile_table(y_true, y_score, change_deciles=10)

For model/dataset workflows, ModelPlotPy prepares comparable plot input and the top-level plot_* functions render response, gain, lift, profit and ROI views.

The score passed to a binary decile analysis should represent the probability or ranking score for the class you intend to treat as positive. A perfectly drawn plot with the wrong class column is still wrong analysis.