Decision Engine: Adding business logic to models
Posted by Assaf Binstock
on 2021-12-25 15:55:29
Introducing a new decision engine feature on the EyeOnRisk to ensure you get the right answers for your organization
Every business is unique. This is the reason you have a team focused on building models that fit your organization: They consider business goals. They consider your business environment and regulations. They consider your customer journey and products. They consider the model management platform you work with. And they must consider your organization’s unique business logic and policies.
Why is business logic crucial in modelling?
Infusing your model with business logic via a “decision engine” allows you to include and exclude specific data. It lets you hard-code rules that will affect the model even when the data changes Sometimes, it’s because the data is too volatile. Other times, it has to do with the organizations’ policy and strategy. Young companies who have yet to acquire enough data to train predictive models based on machine learning can also benefit from rule-based systems/models.
If you own a young company with a need for a full solution - contacts us now.
Automating the decision engine: A new feature on the EyeOnRisk
After years of providing a unique platform for developing predictive models for risk, we are excited to announce a new feature in our EyeOnRisk version 2.0. A built-in Decision Engine provides an easy, intuitive way to develop rule based logic in a no-code environment. The Decision Engine, and the underlying new component called “Rule Set” is a fundamental part of our solution and will allow you to deploy a FULL SOLUTION for scoring.
As a fundamental part of the system, the Rule Sets and Decision Engine is tightly woven into the entire platform. It follows a full flow (or many flows) and is used in deployments. The inputs and outputs for a Rule Set are monitored and registered in the system for easy access and analysis.
If you’ve been waiting for the perfect decision engine - contacts us now.
Technical Introduction to Decision Engine
A Decision Engine is a business logic implementation layer which allows full control on the decision process governing your models. The platform provides a highly customizable decision engine which allows us to build the most accurate business logic for your business.
In the platform, the actual implementation of business logic is done with the design of a component called “Rule Set”. You may have as many Rule Sets as you wish in the system.
A Rule Set is a series of business logic rules which allows you to implement any type of logic required by your use-case. For example, in the credit-risk use case, the Rule Set may implement things like:
- Hard rules (e.g., no loans for under 25 years of age)
- Govern the threshold over the PD
- The allowed loan amount limits
- General loan limits
- And many more
Building a Rule Set
Implementing a Rule Set is easy and done through the Decision Engine module of the system.
Rule Set may have as many statements as you like. Each statement is a stage in the business logic implementation. All the statements in a Rule Set, together, calculate the entire set of decisions needed to be taken for a given record which is run through the system. As other elements in the EyeOnRisk platform, defining Rule Set statements is a breeze. The intuitive UI will walk you through the process and will allow you to create and test your business logic in minutes.
Here you can see the rule setting screen from the platform: This is a statement of a single rule set.
You will usually create a Rule Set in conjunction to a certain model or models. Models are created in the scope of a flow. This way your business logic is “running” and using the result of a previously defined model. A Rule Set can use as many flows as needed. This means you can combine a set of flows developed for various populations into a unified model using the business logic.
Testing a Rule Set
It’s very important to test the Rule Set and see it running using test data before proceeding. The platform provides a convenient way to test the Rule Set in action. You can use an external file and load it, or use an existing flow as source of data or manually provide a record of data points.
The combination of a Decision Engine together with the easiest model development platform out there, gives you a unique opportunity to revolutionize your model development processes. These unique times call for new methods, improved performance and better processes.
Take the EyeOnRisk platform for a trial today. Contact us to schedule a demo.