plotting.observation#

Visualization functions for observation model results.

Functions

`plot_dimensionalities`	Plot the number of each dimension type.
`plot_var_exp`	Plot shared variance explained by each dimension type.
`plot_dims_pairs`	Visualize pairwise dimensionality analysis.
`plot_var_exp_pairs`	Visualize pairwise shared variance analysis.

plot_dimensionalities( num_dim: ndarray, dim_types: ndarray, sem_dim: ndarray | None = None, group_names: list[str] | None = None, plot_zero_dim: bool = False, ax: Axes | None = None, ) → None[source]#

Plot the number of each dimension type.

Parameters:

num_dimndarray of shape (n_dim_types,): Number of each dimension type.
dim_typesndarray of shape (n_groups, n_dim_types): Binary array indicating which groups are involved in each dimension type.
sem_dimndarray or None, default None: Standard error of the mean for each dimension type, shape (n_dim_types,).
group_nameslist of str or None, default None: List of group names for labeling. If None, uses “1”, “2”, etc.
plot_zero_dimbool, default False: Whether to plot dimension types with zero cardinality.
axAxes or None, default None: Axes on which to draw. If None, uses current axes.

Examples

>>> num_dim, _, _, dim_types = model.obs_posterior.compute_dimensionalities()
>>> plot_dimensionalities(num_dim, dim_types)

plot_var_exp( var_exp: ndarray, dim_types: ndarray, sem_var_exp: ndarray | None = None, group_names: list[str] | None = None, plot_zero_dim: bool = False, fig: Figure | None = None, ) → None[source]#

Plot shared variance explained by each dimension type.

Parameters:

var_expndarray of shape (n_groups, n_dim_types): Fraction of shared variance explained by each dimension type in each group.
dim_typesndarray of shape (n_groups, n_dim_types): Binary array indicating which groups are involved in each dimension type.
sem_var_expndarray or None, default None: Standard error of the mean for variance explained, shape (n_groups, n_dim_types).
group_nameslist of str or None, default None: List of group names for labeling. If None, uses “1”, “2”, etc.
plot_zero_dimbool, default False: Whether to plot dimension types with zero cardinality.
figFigure or None, default None: Figure on which to draw. If None, uses current figure.

Examples

>>> _, _, var_exp, dim_types = model.obs_posterior.compute_dimensionalities()
>>> plot_var_exp(var_exp, dim_types)

plot_dims_pairs( pair_dims: ndarray, pairs: ndarray, n_groups: int, sem_pair_dims: ndarray | None = None, group_names: list[str] | None = None, fig: Figure | None = None, ) → None[source]#

Visualize pairwise dimensionality analysis.

Parameters:

pair_dimsndarray of shape (n_pairs, 3): Dimensionalities for each pair: [total_group1, shared, total_group2].
pairsndarray of shape (n_pairs, 2): Indices of groups in each pair.
n_groupsint: Total number of groups.
sem_pair_dimsndarray or None, default None: Standard error of the mean for pairwise dimensionalities, shape (n_pairs, 3).
group_nameslist of str or None, default None: List of group names for labeling. If None, uses “1”, “2”, etc.
figFigure or None, default None: Figure on which to draw. If None, uses current figure.

Examples

>>> from latents.observation import ObsParamsPosterior
>>> num_dim, _, var_exp, dim_types = obs_posterior.compute_dimensionalities()
>>> pair_dims, _, pairs = ObsParamsPosterior.compute_dims_pairs(
...     num_dim, dim_types, var_exp
... )
>>> plot_dims_pairs(pair_dims, pairs, n_groups=len(obs_posterior.y_dims))

plot_var_exp_pairs( pair_var_exp: ndarray, pairs: ndarray, n_groups: int, sem_pair_var_exp: ndarray | None = None, group_names: list[str] | None = None, fig: Figure | None = None, ) → None[source]#

Visualize pairwise shared variance analysis.

Parameters:

pair_var_expndarray of shape (n_pairs, 2): Fraction of shared variance explained for each group in each pair.
pairsndarray of shape (n_pairs, 2): Indices of groups in each pair.
n_groupsint: Total number of groups.
sem_pair_var_expndarray or None, default None: Standard error of the mean for pairwise variance explained, shape (n_pairs, 2).
group_nameslist of str or None, default None: List of group names for labeling. If None, uses “1”, “2”, etc.
figFigure or None, default None: Figure on which to draw. If None, uses current figure.

Examples

>>> from latents.observation import ObsParamsPosterior
>>> num_dim, _, var_exp, dim_types = obs_posterior.compute_dimensionalities()
>>> _, pair_var_exp, pairs = ObsParamsPosterior.compute_dims_pairs(
...     num_dim, dim_types, var_exp
... )
>>> plot_var_exp_pairs(pair_var_exp, pairs, n_groups=len(obs_posterior.y_dims))