Here's a thing nobody likes: You're at the checkout register, getting ready to swipe or chip-read your credit card and pay. You know you have enough money, but your credit card company declines the purchase. That can mean switching to another card or, less happily for the shop you're patronizing, skipping the purchase altogether.
A lot of these have to do with fraud detection — the false positives banks' computers catch when they're not sure you're spending the way you usually do. That could soon be a thing of the past, or at least of much less frequency: Researchers at the Massachusetts Institute of Technology have come up with a machine-learning program to reduce false-positive fraud reports on credit card charges by more than half.
Video of the Day
Right now, it's a worse problem for banks than it is for consumers. Only about 1 in 5 of these false positives are actually fraudulent, which means a lot of money and time down the drain trying to sort it all out. Meanwhile, customers are annoyed and businesses can lose about $118 billion in annual revenue, thanks to purchases that are never completed.
You can read more from MIT if you want specifics on how it all works. Meanwhile, if you think you've run astray of some actual fraudsters, you should use some online tools to double-check. Identity thieves have a certain type, by and large; even if you don't fit the profile, there are lots of government agencies and private companies you can turn to for help.