Friday, 30 September 2016

Podcast: Fraud Detection and Machine Learning

Fraser, Peter, and Nick discuss how machine learning can be used to fight online fraudsters, with special guest Mairtin O'Riada, CIO of Ravelin.



To subscribe to the podcast, add this RSS feed to your preferred player. http://feeds.soundcloud.com/users/soundcloud:users:219479129/sounds.rss

1 comment:

Chris Lear said...

This very much reminds me of spam and virus filtering. Virus filters (as far as I know) have no machine learning component. They just scan for a list of known viruses, that list being updated on a schedule. The most effective spam filters (as far as I know) have very few specific rules, and can tell just by looking at the raw email whether it has a high probability of being spam. Most of the useful information is in the SMTP headers, not in the email text. The training of the machine is done by all the users, who very quickly file unwanted email in their spam folder, so the feedback is super-quick and the corpus is immense. Nobody seems to mind about the lack of explanatory value in the spam filter's Bayesian reasoning.