Generalized cross-covariances and their estimation

摘要: Generalized cross-covariances describe the linear relationships between spatial variables observed at different locations. They are invariant under translation of locations for any intrinsic processes, they determine cokriging predictors without additional assumptions and unique up to functions. If model is stationary, that if variograms bounded, correspond stationary cross-covariances. Under some symmetry condition equal minus usual cross-variogram. We present a method estimate these generalized from data arbitrary sampling In particular we do not require all same points. For fitting coregionalization combine this new with standard algorithm which ensures positive definite matrices. study behavior both by computing variances exactly simulating various models.