hookeai.model_architectures.procedures.convergence_plots.plot_prediction_loss_convergence_uq

plot_prediction_loss_convergence_uq(models_base_dirs, training_dirs, predictions_dirs, filename='testing_loss_convergence_uq', save_dir=None, is_save_fig=False, is_save_plot_data=False, is_stdout_display=False, is_latex=True)

Plot average prediction loss versus training data set size.

Uncertainty quantification data accounting for different model samples predictions for each training data set size is required. The corresponding directory named ‘uncertainty_quantification’ should exist in each model base directory.

Parameters:

• models_base_dirs (tuple[str]) – Base directory where each model is stored.

• training_dirs (tuple[str]) – Directory where each model training data set is stored.

• predictions_dirs (tuple[str]) – Directory where each model samples predictions results files are stored.

• filename (str, default='testing_loss_convergence') – Figure name.

• save_dir (str, default=None) – Directory where figure is saved. If None, then figure is saved in current working directory.

• is_save_fig (bool, default=False) – Save figure.

• is_save_plot_data (bool, default=False) – Save plot data. Plot data is stored in a file with a single dictionary where each item corresponds to a relevant variable used to generate the plot. If the figure directory is provided, then plot data is saved in the same directory, otherwise is saved in the current working directory.

• is_stdout_display (bool, default=False) – True if displaying figure to standard output device, False otherwise.

• is_latex (bool, default=False) – If True, then render all strings in LaTeX. If LaTex is not available, then this option is silently set to False and all input strings are processed to remove $(…)$ enclosure.