4 ms·
"Machine learning" and other techniques have actually been pretty effective at reducing card fraud in recent years -- an overall loss of just 0.07%. And this i
by abalone 13y ago
"Machine learning" and other techniques have actually been pretty effective at reducing card fraud in recent years -- an overall loss of just 0.07%.
And this is all done with a system that is very convenient for consumers to use correctly for everyday use, without significant risk of a total loss due to getting hacked or having their password or other security factors compromised.
Source: http://en.wikipedia.org/wiki/Credit_card_fraud http://en.wikipedia.org/wiki/Credit_card_fraud
- FatalLogic 13y ago>an overall loss of just 0.07%. No, that's only the direct loss for credit card issuers. It's not surprising that the companies that control the system limit their own losses as much as they can, by passing on the costs to others. The losses to merchants, etc are far greater. In fact that figure understates even the card issuers' total losses, because it doesn't include billions of dollars in "fraud solutions expense, IT, staffing, outsourcing, and outside data/investigation expenses" The report from which 0.07% was taken indicates immediate losses that are about 15x higher. In other words, around 1% The report makes it extremely clear that the 0.07% figure is a severe underestimate of the real costs. In fact, that is really the main topic of the report. It's unfortunate that all that research has been ignored and the almost meaningless number has been cherry picked for Wikipedia. Link to the report (PDF) which is the source of the 0.07% figure: http://web.archive.org/web/20091229101826/http://www.sas.com/news/analysts/mercator_fraud_1208.pdf http://web.archive.org/web/20091229101826/http://www.sas.com...