Fraud & Risk
Demystifying Machine Learning
Machine learning is a powerful fraud prevention technology, but it is not a panacea. This presentation examines what machine learning can actually do, with an inside look at classification and decision boundaries as well as common model types and their properties. Guidelines for selecting a framework are also discussed, followed by caveats on automation. The presentation concludes with pros and cons of machine learning and several takeaways a business should evaluate when considering whether machine learning is a good fit for its risk department.