To predict fraudulent claims for term insurance and create a fraud intelligence view.

One of the prominent life insurance company intends to find and predict fraudulent claims for Term Insurance. Major challenge was to understand the variables which are causing the propensity for fraud and use the historical claims data to create a profile of the potential fraudster and initiate proactive investigation further. A smarter UI needs to be designed for Causal factors and drill downs for the same with downloadable reports.

We had performed a data quality assessment for claims, along with the analysis of fraud claims detection and key factor comparison across models. In order to understand the variables tending to fraud, we had studied the impact caused from these variables and created prediction plots, to combine and identify the impact of the factors on each individual. Such an approach had enabled us in generating various fraudster profiles and commence a deeper look-over.

After gaining a clear understanding on all variables and its impact, a Fraud Intelligence View has been created for our client to get practical insights and a holistic view of the evolving landscape. With the help of an interface (UI) created exclusively for model evaluation, our client has got the Fraud Prediction with an accuracy greater than 85%.