Model Build Phase

The Model Build phase walks you through building a highly accurate machine learning models that are specific to the forecasting problem you are trying to solve.

NOTE: You must have a business administrator or higher security role to access pages in the Model Build phase.

TIP: The Consumption Groups page cannot be accessed until the Data > Dataset page has run the job to merge the data sets. The page can always be accessed in the Utilization phase.

Model building in Sensible Machine Learning is a three-section phase where you configure, build, and deploy machine learning or statistical models for time series forecasting in a Sensible Machine Learning project. It begins with sourcing data into Sensible Machine Learning. The data is typically sourced from an external database connection.

From there, you can add various machine learning parameters to your project. This includes parameters such as locations, events, forecast ranges, and model configurations. Most of these are defined in the Configure section.

After specifying the data and the configurations to generate, engineer, and transform the data, use the Model Build phase Pipeline page to run a model pipeline.

IMPORTANT: You should have a thorough understanding of the data in any target or feature data source used in the machine learning or statistical model you are working with. Various steps in the Model Build and Utilization phases let you review and verify the data being used and how the data is configured for Sensible Machine Learning.

NOTE: Times shown are in the specified time zone. However, custom time frames use the actual prediction data generated dates.

See Appendix 2: Use Case Example for information about models used by Sensible Machine Learning.