Enhancing rules with Machine Learning predictions
Churn is an important metric for borrower retention and as such is critical in decision-making processes. In this tutorial, you enhance the Miniloan rule project with borrower churn predictions from Machine Learning.
The credit industry is a competitive landscape. Institutions are vying not only to attract new borrowers but to reduce borrower churn, which is defined as a borrower's propensity to change institutions. But if financially knowledgeable borrowers are willing to change institutions, how you identify borrowers that are likely to churn to provide incentives to stay? Machine Learning predictions can enhance your prescriptive rules with the foresight necessary to anticipate future occurrences.
In this tutorial, you enrich the Miniloan rule project with borrower churn predictions from Machine Learning to identify financially savvy borrowers with a high probability of churn.
Learning objectives
By the end of this tutorial, you complete the following steps that are necessary for integrating Machine Learning predictive models into your business decisions:- Deploying a rule project and a Machine Learning model
- Exploring Machine Learning
- Importing a YAML file into Rule Designer
- Writing rules that invoke a Machine Learning reference
- Writing rules that use Machine Learning predictions
- Adding rule packages to a ruleflow
- Submitting the HBRMINW JCL
Time required
Expect to spend 60 minutes on this tutorial. If you explore other concepts related to this tutorial, it could take longer to complete.Prerequisites
Before you start this tutorial, you must:- Install Rule Designer
- Install Liberty, available in the Operational Decision Manager installation files
- Obtain access to the IBM Watson Machine Learning for IBM z/OS platform and the ChurnModel sample model
You can also review the tutorial Getting started with business rules
to understand the Miniloan
scenario and concepts.