DefinePK

Abstract

In semivariogram modeling, data characteristics greatly influence the steps involved in the semivariogram modeling. Homogeneous data will be better modeled with simple semivariogram models such as: exponential, Gaussian, and spherical models. Meanwhile, heterogeneous data are expected to be better modeled with a mixture semivariogram models. A mixture model is a combination of several simple semivariogram models of a certain proportion. The proportion for each model can be determined from the mean squared error (MSE). If the MSE value is smaller, then the proportion of the corresponding simple models will be greater. Even though the mixture is more complicated, the model can be an alternative in semivariogram modeling which allows to give MSE values that is not much different than MSE values yielded by using the simple models.