Postprocess

Post-process is used to transform model predictions after completion of a simulation. As soon as post-process parameters are changed, all predictions and other simulation results are updated on the fly.

Average of of top-ranked models is used to reduce variance of residuals. However posteriori selection of the number of top-ranked models might lead to over-fitting and lower accuracy of predictions.

Round to nearest integer replaces predicted values with integer numbers.

Replace negative predictions with 0 is used to replace negative predictions with zero value.