decoupler.plot_metrics_boxplot

decoupler.plot_metrics_boxplot(df, metric, groupby=None, figsize=(5, 5), dpi=100, ax=None, title=None, xlabel=None, ylabel=None, return_fig=False, save=None, **kwargs)

Plot boxplots showing the distribution of scores between methods for a metric.

Parameters:
dfDataFrame

Performance metrics per method, obtained by running run_benchmark.

metricstr

Name of metric to plot, must be either “mcauroc”, “mcauprc”, “rank” or “nrank”.

groupbystr

Metrics can be gruped by an extra categorical column.

figsizetuple

Figure size.

dpiint

DPI resolution of figure.

axAxes, None

A matplotlib axes object. If None returns new figure.

titlestr

Text to write as title of the plot.

xlabelstr

Text to write as xlabel of the plot.

ylabelstr

Text to write as ylabel of the plot.

return_figbool

Whether to return a Figure object or not.

savestr, None

Path to where to save the plot. Infer the filetype if ending on {.pdf, .png, .svg}.

kwargsdict

Other keyword arguments are passed through to seaborn.boxplot().

Returns:
figFigure, None

If return_fig, returns Figure object.