structboost.plot_boosting_coefficient_paths

structboost.plot_boosting_coefficient_paths(beta_path, gene_names, cluster_labels, *, top_m=10, linewidth=1.6, alpha=0.9)[source]

Plot coefficient trajectories across boosting steps (per target).

This visualizes how coefficients evolve over iterations. For each target, we select the top-|beta| features at the final step and plot their paths.

Parameters:

  • beta_path (ndarray) – Array of shape (n_targets, n_steps, n_features), typically history.beta_path from allboost(…, return_history=True).

  • gene_names (Sequence[str]) – Feature names of length n_features.

  • cluster_labels (Sequence[str]) – Target labels of length n_targets.

  • top_m (int) – Number of features to display per target (ranked by |beta| at final step).

  • linewidth (float) – Line width for each feature trajectory.

  • alpha (float) – Line alpha for each feature trajectory.

Returns:

figs, axes – Lists of matplotlib Figure/Axes objects, one per target.