Hyperparameter Tuning

Hyperparameters are the algorithm settings under Prediction Algorithm & Settings. The option to change the Number of runs for tuning each set essentially allows you to run multiple simulations to determine the algorithm settings that produce the best results. Again, the defaults will most likely yield the best results, but if you want to increase the number of runs, you can use the calculated algorithm settings for your final computation. Note that the higher the number of runs, the longer the process takes. You can also change the Maximum # of runs which may further refine the results. However, remember that the data research team found that the default of 100 generally produced optimal results.

An online document —Production Prediction Workflow—details the steps for creating a production prediction model using the Zones table, discusses examples of hyperparameter tuning, and also shows how to set optimal engineering constants and grid data to run the model on the Grids table.