Kriging is a moderately quick interpolator that can be exact or smoothed depending on the measurement error model. It is very flexible and allows you to investigate graphs of spatial autocorrelation. Kriging uses statistical models that allow a variety of map outputs including predictions, prediction standard errors, probability, etc. The Flexibility of Kriging can require a lot of decision-making. Kriging assumes the data come from a stationary stochastic process, and some methods assume normally-distributed data.
Note: 1. The process of kriging(prediction): binning the neighborhood points of the unmeasured point - getting the semivariogram values of the neighborhood bins of the unmeasured point from the fitted model - getting the semivariogram values between the unmeasured points with the neighborhood measured points - getting the weights of neighborhood points - getting the prediction value of the unmeasured point.
2. Kriging is divided into two distinct tasks: quantifying the spatial structure(getting the weights ) of the data and producing a prediction(creating surface);
3. The addition of a statistical model, which include probability(a most important feature in statistics), separates kriging methods from the deterministic methods;
4. Kriging model rely on the notion of autocorrelation.