Spatial Analysis Using Big Data

4.4. Kriging

4.4.1. Spatial prediction and Kriging

Since geostatistical models assume weak stationarity or intrinsic stationarity in spatial processes, they can be used for spatial prediction of random variables at arbitrary points in a natural way. The term “prediction,” which is often used in the sense of predicting the future, can be misleading, but in this context, it designates the estimation of random variables at arbitrary points other than the observation points. In this document, “spatial prediction” and “prediction” are used without distinction.

Due to budget or technology-related constraints, or even privacy protection policies, spatial data can be measured at only limited observation points. For this reason, in many cases, it ...