Among all the online purchases that occur every day, a few are frauds. Finding these is a hard and cumbersome task and today we use a semi-automatic system for this. The next phase is to improve fraud detection through machine learning. We present the process of converting raw data into features, training models on these features and getting a prediction system up and running.
Staffan Ekvall received his Ph.D. in Machine Learning from the KTH Royal Institute of Technology in 2007. After spending a few years in various fields he joined Svea Ekonomi in 2011 and has played an important role in building their online payment platform and anti-fraud system. His interests are focused on applied machine learning for automatic decision making, but he is also interested in deep learning and AI in general.
Gunnar Dahlberg is a senior systems developer at Svea Ekonomi working with the online payment platform and anti-fraud system. He received a M.Sc. in Bioinformatics Engineering from Uppsala University in 2010. Before joining Svea Ekonomi, Gunnar worked as a consultant and developer for various companies. His has a special interest in developing data warehouses, ETL-processes and building applications based on machine learning and applied statistics.
Brought to you by: