loading

Business Impact

  • Identification of policies with the highest claim probability

  • Determining the factors that may lead to a policy ending up with a claim

  • 84%

    Model Accuracy

Customer Key Facts

  • Location : North America
  • Industry : Insurance

Problem Context

The customer, a global insurance and reinsurance provider, was looking for a single metric that will help them assess the risk associated while writing a policy. They wanted to determine the factors that may lead to a policy ending up with a claim and obtain insights into the impact of a combination of policy characteristics.

Challenges

 

  • Deciding the granularity of claim prediction as data rolled up at policy level was less feature-rich
  • Finalizing a single metric to adjudicate underwriting of policy
  • Need of a real-time prediction engine for better decision making

Technologies

Apache Zeppelin
dmlc XGBoost
R studio

Solution

Quantiphi developed a ML model for the Prediction of Loss likelihood associated with a policy to assist underwriters

Result

  • Identification of customers with high risk of claims
  • Improved loss reserving
  • Client was able to undertake actions to mitigate the risk of claims

Looking for similar project?

Let's Talk

Get your digital transformation started

Let's Talk