GaussPrior

phasic.svgd.GaussPrior(mean=None, std=None, ci=None, prob=0.95)

Gaussian prior distribution.

The prior is defined in THETA space (the natural parameter space). When used with positive_params=True, SVGD automatically handles the transformation to PHI space with proper Jacobian correction.

Can be specified via mean/std or credible interval.

Parameters

mean : float = None

Prior mean in THETA space. Required if std is provided.

std : float = None

Prior standard deviation in THETA space. Required if mean is provided.

ci : tuple of (float, float) = None

Credible interval (low, high) in THETA space. Alternative to mean/std.

prob : float = 0.95

Probability mass in the credible interval (only used with ci).

Examples

>>> # Specify via mean and std
>>> prior = GaussPrior(mean=5.0, std=2.0)
>>>
>>> # Specify via 95% credible interval
>>> prior = GaussPrior(ci=(2.0, 8.0))
>>>
>>> # Plot to verify prior matches your beliefs
>>> prior.plot()  # Shows Gaussian centered at 5
>>>
>>> # Use in SVGD - transformations handled automatically
>>> svgd = graph.svgd(data, theta_dim=1, prior=prior)

Methods

Name	Description
plot	Plot the Gaussian prior distribution in THETA space.
sample	Sample from the prior.

plot

phasic.svgd.GaussPrior.plot(log=False, ax=None, return_ax=False, **kwargs)

Plot the Gaussian prior distribution in THETA space.

Parameters

log : bool = False

If True, plot log-probability instead of probability density.

ax : matplotlib.axes.Axes = None

Axes to plot on. If None, creates new figure.

return_ax : bool = True

If True, return ax. If False, call plt.show() instead.

****kwargs** : = {}

Additional arguments passed to plot function.

Returns

: matplotlib.axes.Axes

The axes with the plot (only if return_ax=False)

sample

phasic.svgd.GaussPrior.sample(key, shape)

Sample from the prior.

When _transform is set, samples in THETA space and converts to PHI space.

Parameters

key : jax.random.PRNGKey

Random key

shape : tuple

Shape of samples (n_particles, theta_dim)

Returns

: array

Samples (in PHI space if transform is set)