Blog

Risk Ident at MRC Seville

Risk Ident at MRC Seville

The Merchant Risk Council provides several occasions every year where companies can exchange views on payments and fraud in ecommerce. The Risk Ident team was out in force at MRC Seville and we were asked to share our expert insights on one of the biggest current topics: machine learning.

For the payments and fraud prevention industry, MRC Seville(18. – 20. May) was a very important event. Over 450 attendees took the chance to expand their network, listen to presentations and interact in panel discussions. At the Risk Ident booth, companies had the chance to learn about our products and learn important insights into the newest approaches to fraud prevention.

Dr. Mario Elstner, Tech Lead Data Science at Risk Ident, had the pleasure of presenting to attendees on “Demystifying machine learning”, in which he showed how Risk Ident is applying machine learning to fight online fraud. His message: machine learning is a very powerful tool in improving fraud prevention, but it is not the silver bullet that many people had hoped would solve fraud once and for all.

At Risk Ident we know that to get great results it is not enough to put data into some kind of supercomputer and press ‘start’. What you need is experts with vast domain knowledge, which they gained from years of fighting against fraud. They know how fraudsters act and they know how to detect new fraud patterns.

In fraud prevention, the quality of data is of great importance – and it is even more important to refine the data. This is called “feature engineering”, a process of using domain knowledge of the data to create features that make machine learning algorithms work. It is easy to see that you have to put the right features into the machine to get good results out of it.

But data refinement is a job for experts – human experts with domain knowledge. With their experience they know how to handle different fraud scenarios for different companies and even different industries. Fully-automated machine learning out of the box sounds great, but if you truly want to reduce fraud, it’s more of a myth; the reality is that man and machine − together − are our best defence.