acfr/eup3m: Evaluating the Uncertainty and Predictive Performance of Probabilistic Models - EUP3M v1.0.1

Features

• This research repository demonstrates concepts in probabilistic prediction and model validation particularly in relation to ore grade estimation in mineral deposits.
• The main components enable model construction and performance evaluation
• Modelling approaches include kriging and Gaussian process, with/without sequential and correlated random field simulations
• Statistical analysis focuses on global accuracy, spatial fidelity and uncertainty aspects of predictive distributions

Changes

• Renamed attributes consistent with the description in the updated docs/background.pdf document
• Fixed bug in _downsample_worker