Model Build Phase Pipeline Section

The Pipeline section of the Model Build phase is where you run the Sensible Machine Learning pipeline. This brings all prior configurations and parameter selections together. The pipeline:

• Generates, transforms, and selects features based on predictive capability.

• Selects optimal hyperparameter sets for each model of each target.

• Trains models and tests them on historical data.

Running a pipeline produces the best model for each given target. After running a pipeline, you can view various statistics and metrics before deploying the models for utilization.

In the Pipeline section, the XperiFlow engine uses all the information and configurations gathered from the pages in the Data and Configuration sections to run feature engineering for each target.

While creating numerous new features, the engine also iteratively selects the best features to keep for each target. These are later used for all the models run for each specific target.