Speakers Bio & Abstract

Dennis Weinbender Global Risk Management Department

Dennis Weinbender joined the Global Risk Management Department of Commerzbank in Frankfurt in 2011. Since then he has been implementing new models for early risk detection, most of which are still in active use. After several stops in Hamburg, Amsterdam and Frankfurt, he decided to focus on applications of machine learning and fraud detection in 2013. After implementing the current fraud detection system for retail business, he has worked on modeling fraud patterns in financial information for small, middle and large corporations.Abstract
Using Geospatial Data and Machine Learning to Detect Fraud Patterns.
Fraudulent entities are corporates with a very high risk profile, masking their real situation using information policies. In these cases, after very high short term profitability a complete loss of the bank exposure is realized at default. Therefore, fraud prevention is not only a regulatory issue but also an economic necessity. In this talk on detecting fraudulent corporate entities, we present a methodology built and tested using information from more than 1,000,000 corporate entities and 10 000 000 private customers. Geospatial patterns are one of the main sources of information. They enable us to link entities into networks and extract superior signals using state-of-the-art machine learning methodologies.