kld_gaussian

This function computes \(D_{KL}(p||q)\), where \(p\sim \mathcal{N}(\mu_1,\Sigma_1)\)
and \(q\sim \mathcal{N}(\mu_2,\Sigma_2)\).

Estimation algorithms for Kullback-Leibler divergence between two probability
distributions, based on one or two samples, and including uncertainty quantification.
Distributions can be uni- or multivariate and continuous, discrete or mixed.

Niklas Hartung

kldest

Sample-Based Estimation of Kullback-Leibler Divergence

kld_gaussian function

<dl><dt>mu1</dt>
<dd>A numeric vector (mean of true Gaussian)</dd>
<dt>sigma1</dt>
<dd>A s.p.d. matrix (Covariance matrix of true Gaussian)</dd>
<dt>mu2</dt>
<dd>A numeric vector (mean of approximate Gaussian)</dd>
<dt>sigma2</dt>
<dd>A s.p.d. matrix (Covariance matrix of approximate Gaussian)</dd></dl>

Arguments

Analytical KL divergence for two uni- or multivariate Gaussian distributions — kld_gaussian

<dl>

<dt>mu1</dt>
<dd>A numeric vector (mean of true Gaussian)</dd>


<dt>sigma1</dt>
<dd>A s.p.d. matrix (Covariance matrix of true Gaussian)</dd>


<dt>mu2</dt>
<dd>A numeric vector (mean of approximate Gaussian)</dd>


<dt>sigma2</dt>
<dd>A s.p.d. matrix (Covariance matrix of approximate Gaussian)</dd>

</dl>

kld_gaussian: Analytical KL divergence for two uni- or multivariate Gaussian distributions

Description

Usage

Value

Arguments

Examples