Research Project: RATE-Analytics: Next generation predictive analytics for data-driven banking and insurance
Description
In this project co-funded by NWO Commit2Data, Rabobank and Achmea, we developed foundations and techniques for advanced analytics with an expert in the loop. We have proposed several novel techniques bringing expertise in predictive analytics and visual analytics. We facilitated construction of more effective and interpretable data-driven solutions applicable in a wide range of use cases in banking and insurance. We demonstrated their potential on real world datasets and conducted several use cases in collaboration with Achmea and Rabobank. Notably, we developed explainable machine learning techniques for providing explanations for fraud prediction in sick leave insurances, anomaly detection in payment transaction. We also explored applicability of different state of the art approaches for predicting customer churn for checking accounts, and fraud detection and mortgage prepayments prediction.
TU/e team included, dr. Samaneh Khoshrou, dr. Dennis Collaris, prof. Jack van Wijk, and prof. Mykola Pechenizkiy
Details
- Principal Investigator
-
Mykola Pechenizkiy
- Involved members: