An Embedded Model Estimator for Non-Stationary Random Functions Using Multiple Secondary Variables

Abstract An algorithm for non-stationary spatial modelling using multiple secondary variables is developed herein, which combines geostatistics with quantile random forests to provide a new interpolation and stochastic simulation. This paper introduces the method shows that its results are consistent similar in nature those applying geostatistical forests. The allows embedding of simpler techniques, such as kriging, further condition model. works by estimating conditional distribution target variable at each location. family distributions called envelope variable. From this, it possible obtain estimates, quantiles uncertainty. also produce simulations from envelope. As they sample envelope, realizations therefore locally influenced relative changes importance variables, trends variability.