A Comparison between the Performance of kriging and cokriging in Spatial Estimation with Application

This paper deals with application of spatial prediction techniques (universal kriging and cokriging) to predict unmeasured locations in mining field such as mineral ores. However, the combination of the kriging and cokriging as many predictive techniques is still an active research area to obtain an adequate prediction model. The aim is to obtain solution of spatial prediction using multivariate. We experiment primary and secondary variables of the two techniques to create a prediction model that correlated covariance functions. Practically, we apply the model on real data samples of (120) of Copper and Nickel metals that is taken from the Korf property. We are able to minimize the error rates and satisfy the weights constraints comparing with Gaussian and Power models.