Struct MatrixNormal 

pub struct MatrixNormal { /* private fields */ }

Matrix normal distribution over real matrices.

Implementations

impl MatrixNormal

pub fn new( mean: DMatrix<f64>, row_cov: DMatrix<f64>, col_cov: DMatrix<f64>, ) -> Result<MatrixNormal, String>

Construct a matrix normal distribution with mean matrix mean, row covariance row_cov, and column covariance col_cov.

Errors

Returns Err if covariance dimensions do not match the mean or either covariance is not symmetric positive-definite.

Trait Implementations

impl Display for MatrixNormal

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<R: Rng + ?Sized> Distribution<R> for MatrixNormal

type Domain = Matrix<f64, Dyn, Dyn, VecStorage<f64, Dyn, Dyn>>

The type of values produced by this distribution.

fn sample(&self, rng: &mut R) -> DMatrix<f64>

Draw one random sample from this distribution.

fn log_prob(&self, x: &DMatrix<f64>) -> f64

Compute the log probability (discrete) or log probability density (continuous) of observing x under this distribution.

fn is_discrete(&self) -> bool

Returns true if this distribution is discrete (e.g. Bernoulli, Poisson), or false if it is continuous (e.g. Normal).

fn log_cum_prob(&self, _x: &Self::Domain) -> f64

Compute log P(X <= x) for discrete distributions, or the log CDF for continuous distributions.