Everywhere you look today, there are examples cropping up of how machine learning is revolutionising different industries. In media and entertainment, Spotify and Netflix sort through billions of data points to find patterns in music, films and television that consumers have enjoyed — and then make suggestions based on their tastes. In retail, Amazon prompts consumers to buy everything from nappies to office chairs based on shoppers’ previous purchases. In finance, machine learning is helping investors anticipate market trends and powering innovations underlying everything from self-driving cars to voice-assistant applications.
In the payments industry, machine learning is similarly becoming an increasingly important tool to help businesses combat fraud. As new technologies transform the way we pay — originally credit and debit cards and, more recently, with kiosks, smartphones and other mobile devices — the number of transactions flowing through global payment networks has increased. At the same time, criminals have grown more sophisticated and more adept at using technology — even big data and analytics — to disguise illicit activity. As a result, it is getting harder and harder for businesses to determine which transactions to approve and which ones to reject.
Machine learning methodologies, when deployed as part of automated fraud screening systems, can help businesses make the right call. At CyberSource, we have long understood the power of machine learning, and it has underpinned our fraud management solution for the last 15 years.
The earliest version of the CyberSource Decision Manager platform was underpinned by two main elements that define any machine learning system. `To learn more, read the Whitepaper, `Fraud Management – A Machine Learning Approach.’
What is machine learning?
Machine learning relies on complex statistical methods and high-octane computing power. At its core, however, is a very simple concept. By identifying the most influential cause-and-effect relationships from the past, a machine can learn to make accurate predictions about the future.
There are a number of strengths that make machine learning such a powerful approach, specifically when it comes to fighting fraud. Machine learning helps:
Facilitate real-time decision-making. Rules-based systems, where people create ad-hoc rules to determine which types of orders to accept or reject, require a great deal of time-consuming, manual interaction. Machine learning can help evaluate huge numbers of transactions in real time
Improve accuracy. As criminals have grown more sophisticated, they have become more adept at disguising fraud. Machine learning can often be more effective than humans at detecting subtle or non-intuitive patterns to help identify fraudulent transactions. It can also help avoid false positives — good orders that are erroneously identified as fraudulent.
Rapidly respond to change. Because fraudsters are always changing their tactics, it’s a constant cat-and-mouse game. Machine learning is continuously analysing and processing new data and then autonomously updating its models to reflect the latest trends.
Lower costs. Significant advances in technology have reduced the costs associated with machine learning solutions and computing systems capable of running them. As machine learning improves accuracy, it reduces costly false positives and minimises the time and expense of manual reviews.
Learn more
You have been reading an edited excerpt from our whitepaper, Fraud Management: A Machine Learning Approach. To dive into the detail on:
The various machine learning methods, like regression analysis, artificial neural networks and decision trees
How real-time fusion modelling can help businesses more effectively and efficiently manage and detect fraud
Why better data means better results
CyberSource’s long history as a machine learning pioneer, and how it has made Decision Manager the comprehensive, robust platform it is today
... and much more, you can download our whitepaper, Fraud Management: A Machine Learning Approach, here.
About Scott Boding
Scott Boding is a Senior Director, Risk Solutions Product Management at CyberSource. He’s playing a significant role in CyberSources growth as an industry leader in fighting online fraud, being an expert in anti-fraud strategies and rule making.
About CyberSource
CyberSource, a wholly-owned subsidiary of Visa Inc., is a payment management company. Over 400,000 businesses worldwide use CyberSource and Authorize.Net brand solutions to process online payments, streamline fraud management, and simplify payment security. The company is headquartered in Foster City, CA and maintains offices throughout the world, with regional headquarters in Singapore, Tokyo, Miami, Sao Paulo and Reading, U.K. CyberSource operates in Europe under agreement with Visa Europe.
Every day we send out a free e-mail with the most important headlines of the last 24 hours.
Subscribe now