pymc.dims.MvNormal

class pymc.dims.MvNormal(name, *dist_params, dims=None, initval=None, observed=None, total_size=None, transform=UNSET, default_transform=UNSET, model=None, **kwargs)

Multivariate Normal distribution.

Parameters:

muxtensor_like

Mean vector of the distribution.

covxtensor_like, optional

Covariance matrix of the distribution. Only one of cov or chol must be provided.

cholxtensor_like, optional

Cholesky decomposition of the covariance matrix. only one of cov or chol must be provided.

lowerbool, default True

If True, the Cholesky decomposition is assumed to be lower triangular. If False, it is assumed to be upper triangular.

core_dims: Sequence of string

Sequence of two strings representing the core dimensions of the distribution. The two dimensions must be present in cov or chol, and exactly one must also be present in mu.

**kwargs

Additional keyword arguments used to define the distribution.

Returns:

XTensorVariable

An xtensor variable representing the multivariate normal distribution. The output contains the core dimension that is shared between mu and cov or chol.

Methods

MvNormal.dist(mu[, cov, chol, lower, core_dims])