mtopic.pl.select_n_topics

mtopic.pl.select_n_topics#

mtopic.pl.select_n_topics(mdata, *, plateau_threshold=0.05, figsize=(6, 5), transparent=False, save=None)#

Fit a rational function to the mean held-out log-likelihood curve and identify the optimal number of topics.

Requires mtopic.tl.estimate_topic_number to have been run first, which stores the CV results in mdata.uns["CV_results"].

Parameters:

  • mdata (muon.MuData) – MuData object containing mdata.uns["CV_results"] from mtopic.tl.estimate_topic_number.

  • plateau_threshold (float, optional) – Fraction of max slope used to define the plateau onset. Default is 0.05.

  • figsize (tuple, optional) – Figure size in inches (width, height). Default is (6, 5).

  • transparent (bool, optional) – Whether to save the figure with a transparent background. Default is False.

  • save (str, optional) – Path to save the figure. If None, the figure is shown but not saved. Default is None.

Returns:

None. Fit info is stored in mdata.uns["CV_fit_info"].

Return type:

None