A data object summarizes the parameters of the same type or the same business attribute into a business data type, which is usually represented as a JSON object. Reasonable induction is conducive to a better understanding of different business attributes and business parameter descriptions.

We have made several types of constraint recommendations for the data object parameter:

  • required: the data object is invalid if it does not exist or if the Parameter is empty
  • conditionally required: required exists under certain conditions
  • Strong recommendation: if it exists, it will significantly improve decision accuracy, but it is not required
  • Suggestion: if it exists, it will help to improve the decision accuracy but is not required

Data objects can be referenced, subject to the following constraints:

  • If the Parameter referenced by the data object is a "recommended" or "strongly recommended" constraint, then all parameters inside the data object will be changed to a "recommended" or "strongly recommended" constraint regardless of whether there is a "required" or not or "conditionally required" constraints
  • If the Parameter referenced by the data object is a constraint condition of "required" or "conditionally required," then all Parameter inside the data object still follow their respective constraints and ignore external constraints