ggml_ot.pl.contour_hyperparams

ggml_ot.pl.contour_hyperparams(results_df, *, x='alpha', y='reg', fixed_params=None, value_col=('knn', 'mean'), log_axis=True, label_style='package', levels=20, cmap='RdBu_r', pad=True, show=None, save=None)

Contour plot of grid-search results over two hyperparameters.

Parameters:

results_df pandas.DataFrame

DataFrame produced by ggml_ot.tune().

x str, default "alpha"

Hyperparameter for the horizontal axis.

y str, default "reg"

Hyperparameter for the vertical axis.

fixed_params dict or None, default None

Fix remaining hyperparameters to specific values. Keys with None values are averaged across.

value_col str or tuple, default ("knn", "mean")

Column selector for the metric to display.

log_axis bool or "x" or "y", default True

Log10-transform axes (tick labels show original values).

label_style "package" or "paper", default "package"

"package" uses raw parameter names (alpha, reg, …). "paper" uses LaTeX notation (α, λ_F, λ_MI, d).

levels int, default 20

Number of contour levels.

cmap str, default "RdBu_r"

Matplotlib colourmap name.

pad bool, default True

Pad the contour grid to avoid clipping at the edges.

show bool or None, default None

Whether to display the plot. None (default) automatically shows in interactive environments (notebooks, IPython) and suppresses in scripts. True/False override explicitly.

save str, bool, or None, default None

Whether to save the figure to disk. None/False skip saving. True saves to settings.figdir/contour_hyperparams.<figformat>; a str overrides the filename.

Returns:

tuple[matplotlib.figure.Figure, matplotlib.axes.Axes] The figure and axes containing the contour plot.