AutoML Model Optimization

  • Updated

AutoML models optionally automate the grid search step of the model creation process.

AutoML comes with three preconfigured Grid Levels: Low, Medium, and Exhaustive. These settings modify the number of hyperparameter experiment models which will be generated in the grid search. AutoML uses a Bayesian optimizer to reduce the actual grid search space explored. 


By default, AutoML is set to the Medium grid complexity setting, with an optimization approach that attempts to maximize the AUC metric. If you select Auto-Complete from a step prior to the Optimize Model stage, it will process using the Medium grid setting while attempting to maximize the Area Under the Curve of the Receiver Operator Characteristic (AUC). 

The model’s optimization metric can be modified by selecting from the Optimization dropdown. This dropdown will be appropriately labeled as Binomial Optimization or Multinomial Optimization based on the data specification.

Grid settings can be modified by selecting the Edit Initial Parameters slider and modifying the exposed values.


For additional information on hyperparameter settings, see the article on simClassify+ Model Specifications.

When satisfied with the hyperparameter configurations, which will be explored in the grid experiment, select the Next button from the AutoML Navigation Bar. You will be brought to the AutoML Project Info page while the grid executes. While the grid is running, the Project Information page provides users with the status of the grid experiment, presenting the overall progress as well as the current top-performing model (based on the metric selected in the Optimize Model step). 


Other articles on AutoML: 


Was this article helpful?

0 out of 0 found this helpful



Please sign in to leave a comment.