Talk by Märten Veskimäe, Lead Data Scientist @Monese is about about detecting illegal financial transactions. Different modelling approaches are compared that range from linear models to deep learning on graphs. "On the conceptual level, transaction security is a matter of significant importance. Not only because illegal transactions cost trillions of dollars a year globally, but also because a share of that money is linked to serious crime types such as organized crime, human trafficking, terrorism, etc. From a modelling perspective, the topic is related to imbalanced data problems, semi-supervised learning, and it can be approached using a variety of data formats, whether tabular, sequential, or graph data," says Märten Veskimäe
Follow North Star AI on:
Twitter:
Facebook:
0 Comments