Overview
Guardrails ensure the ML App's decisions follow your business rules. You may use one or both of the following types of guardrails to define your business rules:
This article defines each of these types of guardrails and provides more information about how to utilize them.
Fixed Guardrails
Fixed Guardrails compare your ML App's features with a specific text or number. Your ML App's features may be an Influencer, Predicted Goal, or—if you are designing an Advanced ML App—Attribute.
Relative Guardrails
Relative Guardrails compare your ML App's features with one another. Your ML App's features may be an Influencer, Predicted Goal, or—if you are designing an Advanced ML App—Attribute.