🛑Decision model is not about #meachinelearning #ai #llms #data #algorithms #businessrules

Decision modeling is a technique that represents a holistic view of a complex business decision. A business decision is complex and requires multiple techniques, such as #businessrules #meachinlearning #data #optimization and etc. to work together to create an outcome. Therefore, an approach that is based on any of the techniques you use from data science, data and analytics or business rules approach is narrow to that their own specific problems.

✨The Decision-Centric Approach allows you to build the decision model using the decomposition technique. It is a very simple process:
* Decompose a complex business decision to its smaller, more easy-to-understand sub-components (decision units)
* Create relationships between them by looking at their dependencies
* Specify their required inputs for each decision unit
* Specify their metrics based on the KPI of business decision that matters for business

By following the above👆 process, you create a model called a Decision Graph. The decision graph represents a hierarchical, multi-step model for a complex business decision.

Now, for each of the decision units in the graph, you can use different techniques to satisfy its logic implementation.

💣Based on a specific decision unit, the graph:
* Select the appropriate algorithm by considering its mature probabilistic vs deterministic vs predictive …
* If deterministic you can then use #businessrules modeling such as #decisiontable #naturallanguage to implement them
* If they are non-rule logic you can use #tree or #boxedexpression from Decision Model and Notation
* If predictive, you can use the right algorithm based on the specific of the decision unit
* Then select the data required for the algorithm only for that specific decision unit

🚀As you can see in the above approach, just by focusing on the decision unit, your problem scope gets smaller; as the result
* It is easier to solve
* Data is less biased and more focus
* The algorithms become more accurate
* The outcomes are relevant and use multiple techniques

Learn more about the decision graph at What is a Decision Graph?

#DecisionIntelligence

Posted here.

Published July 15th, 2024 at 07:30 am