Business Rule Mining is The Key to the Integrity of Continuous Decisioning
Most companies use domain experts and operations teams in decision automation scenarios to model their business rules. However, anomalies can arise in operating systems whereby it is essential to extract, analyze, and understand the function of underlying business rules, especially when these pertain to customer or market data. This is where the ability to ‘mine’ those rules becomes so important.
By definition, business rule mining is the capability of a system to extract rules from data. This is where a decision automation platform may be used to analyze the data and consequently extract the rules from it.
Business rule mining extracts historical data from databases and systems, and feeds these to the business rule mining module. A decision platform then processes the data and automatically applies an array of Machine Learning algorithms. An outcome is a readable form of existing business rules, such as those shown below:
if (Outlook == ‘sunny’) and (Humidity <= 70) then Play = ‘yes’
if (Outlook == ‘sunny’) and (Humidity > 70) then Play = ‘no’
if (Outlook == ‘overcast’) then Play = ‘yes’
if (Outlook == ‘rain’) and (Windy == false) then Play = ‘yes’
if (Outlook == ‘rain’) and (Windy == true) then Play = ‘no’
Simulate, Debug, Deploy and Explain on Demand
You can then prepare and model a Decision Table and Fact Concept related to those rules automatically on a decision automation platform. These can then be added to the project.
At this point, the user can initiate debugging and simulation activities, and manipulate the rules by feeding them with relevant values for inputs (conditions), as well as retrieving the results (outputs):
The screenshots below demonstrate that after the execution of business rules is complete, the result is set to “Yes”:
A key advantage of this approach is that the user now has full visibility into why a particular decision is being made and is able to explain the logic behind that decision accordingly. In other words, the decision is now seen to be driven by rules rather than a specific machine learning trained model:
The Continuous Decisioning Process
Business rule mining is a crucial part of the continuous decisioning process. FlexRule’s business rule mining module enables even a novice user to utilize machine learning in order to extract business rules from your data with a couple of clicks.
Building this approach into any continuous integration process, such as a CICD pipeline, and subsequently delivering a continuously operating decisioning platform is achieved by looking at the latest dataset in systems, extracting the business rules, building the decision model before testing, debugging, approving, and ultimately going to production. Here is how the continuous decisioning cycle looks:
This approach allows you to constantly monitor and measure the Decision KPI, as well as make any necessary improvements in an iterative and incremental manner.
Learn more here: https://www.flexrule.com/archives/business-rule-mining/
