Normal
The standard parametrisation of the Normal distribution is given by
\[\mathcal{N}(x ; \mu , \Sigma ),\]
where $\mu$ is the mean vector and $\Sigma$ is the covariance matrix. The exact expression for the probabiltiy density function depends on whether $x$ is vector with real or complex values, both are supported. For real valued vectors the density function is given by
\[\mathcal{N}(x ; \mu , \Sigma ) = |2\pi \Sigma|^{-1/2} \exp \Big( -\frac{1}{2}(x-\mu)^* \Sigma^{-1} (x-\mu) \Big),\]
whereas for complex valued vectors the density function is given by
\[\mathcal{N}(x ; \mu , \Sigma ) = |\pi \Sigma|^{-1} \exp \Big( -(x-\mu)^* \Sigma^{-1} (x-\mu) \Big).\]
Types
MarkovKernels.AbstractNormal
— TypeAbstractNormal{T<:Number}
Abstract type for representing Normal distributed random vectors taking values in T.
MarkovKernels.Normal
— TypeNormal{T,U,V}
Standard mean vector / covariance matrix parametrisation of the normal distribution with element type T.
Constructors
MarkovKernels.Normal
— MethodNormal(μ::AbstractVector, Σ::CovarianceParameter)
Creates a Normal distribution with mean vector μ and covariance matrix parametrised by Σ.
MarkovKernels.Normal
— MethodNormal(μ::AbstractVector, Σ::AbstractMatrix)
Creates a Normal distribution with mean vector μ and covariance matrix Σ if Σ is Symmetric / Hermitian. Throws domain error otherwise.
MarkovKernels.Normal
— MethodNormal{T}(N::Normal{U,V,W})
Computes a Normal distribution of eltype T from the Normal distribution N if T and U are compatible. That is T and U must both be Real or both be Complex.
Basics
MarkovKernels.dim
— Methoddim(N::AbstractNormal)
Returns the dimension of the Normal distribution N.
Statistics.mean
— Methodmean(N::AbstractNormal)
Computes the mean vector of the Normal distribution N.
Statistics.cov
— Methodcov(N::AbstractNormal)
Computes the covariance matrix of the Normal distribution N.
MarkovKernels.covp
— Methodcovp(N::AbstractNormal)
Returns the internal representation of the covariance matrix of the Normal distribution N. For computing the actual covariance matrix, use cov.
Statistics.var
— Methodvar(N::AbstractNormal)
Computes the vector of marginal variances of the Normal distribution N.
Statistics.std
— Methodstd(N::AbstractNormal)
Computes the vector of marginal standard deviations of the Normal distribution N.
Probability density function
MarkovKernels.residual
— Methodresidual(N::AbstractNormal, x::AbstractVector)
Computes the whitened residual associated with the Normal distribution N and observed vector x.
MarkovKernels.logpdf
— Methodlogpdf(N::AbstractNormal,x)
Computes the logarithm of the probability density function of the Normal distribution N evaluated at x.
Information theory
MarkovKernels.entropy
— Methodentropy(N::AbstractNormal)
Computes the entropy of the Normal distribution N.
MarkovKernels.kldivergence
— Methodkldivergence(N1::AbstractNormal, N2::AbstractNormal)
Computes the Kullback-Leibler divergence between the Normal distributions N1 and N2.
Sampling
Base.rand
— Methodrand(RNG::AbstractRNG, N::AbstractNormal)
Computes a random vector distributed according to the Normal distribution N using the random number generator RNG.