In ‘big data’ era, making optimal spatial prediction(kriging) can be challenging. Motivated by the fact that traditional method involves inverting n × n covariance matrix which is computational expensive and problematic when n is very large, we study a flexible family of non-stationary covariance functions is defined by using a set of basis functions that is fixed in number r << n , which leads to significant time reduction spatial prediction method that we call fixed rank kriging (FRK) Cressie and Johannesson, 2008. We applied FRK to make prediction on a synthetic CO2 data and meuse data, which includes hundreds of thousands observations.
- Keyword : Best linear unbiased predictor; Covariance function; Frobenius norm; Geostatistics; Mean-squared prediction error; Non-stationarity; Remote sensing; Spatial prediction;
