Predictive and Specialized Modeling • Formula Depot • Overview of the Formula Depot Platform
Previous
 • 
Next
Overview of the Formula Depot Platform
The Formula Depot is a repository to organize, compare, profile, and score models for deployment. Models are prediction formulas. Prediction formulas are saved to the Formula Depot as column scripts. You can add prediction formulas to JMP tables to score data. You can also use the Formula Depot to generate scoring code for prediction formulas to facilitate the deployment of models in environments outside of JMP.
The Formula Depot enables you to perform the following tasks:
•
save prediction formulas outside of data tables
•
save prediction formulas from multiple data tables in a common location
•
save intermediate cleaning formulas within a prediction formula
•
compare models
•
profile models
•
add prediction formulas to a data table for scoring of new data within JMP
•
generate scoring code (C, Python, Java Script, SAS, or SQL) for deploying models outside of JMP
•
The Formula Depot includes intermediate formulas in the generated scoring code. A model constructed from inputs with column formulas includes the corresponding formulas in the scoring code. For example, if your model includes a recoded variable the recode formula is included in the scoring code. This enables you to apply your prediction formula to raw data that has not been processed.
Note: Informative missing is not a supported intermediate formula.

Previous
 • 
Next
Help created on 7/12/2018