normEntropy function

Theoretical value of the entropy for the multivariate gaussian

This function computes the entropy $E_{f} [\log(f)]$ of the density of the multivariate gaussian, with parameters in a list, as it is the case for MCMC target density parameters. This function is used mostly for benchmarking entropy estimation performed by the package (using, e.g., the iid algorithm IID_norm).

normEntropy(target_param)

Arguments

• target_param: A list of two elements: the mean target_param$mean

  and the covariance matrix target_param$v.

Returns

The entropy of the Gaussian with these parameters.

Author(s)

Didier Chauveau.

Examples

d=2 # model dimension
mu=rep(0,d); v = diag(d) # mean and variance
target_param = list(mean=mu, v=v) # parameters
normEntropy(target_param) # the entropy