Entropy and Information Content of Geostatistical Models

摘要: Geostatistical models quantify spatial relations between model parameters and can be used to estimate simulate properties away from known observations. The underlying statistical model, quantified through a joint probability density, most often consists of both an assumed the specific choice algorithm, including tuning controlling algorithm. Here, theory is developed that allows one compute entropy multivariate density when sampled using sequential simulation. self-information single realization computed as sum conditional self-information. average obtained for many independent realizations. For discrete mass functions, measure effective number free parameters, implied by function, proposed. Through few examples, information content related different choices simulation algorithms parameters.