Kriging

Summary

Kriging is a method to predict the value of a random field at an unobserved location from its values at observed locations. The prediction is the best linear unbiased predictor: the weighted average of the data with the smallest mean squared prediction error. The weights come from the spatial dependence of the field, described by its variogram or its covariance function. It is named after the mining engineer Danie Krige, and in the language of Gaussian processes it is essentially Gaussian process regression.