▲ ▼ Online fraud detection systems are broken
Online infrastructure service providers especially the cloud providers ban the customer accounts at whim often due to flawed, outdated fraud detection systems.
There are numerous horror stories[1][2][3] from customers on how their accounts were banned for no apparent reason without proper recourse.
There are genuine cases of fraudulent activities happening using the cloud infrastructure, but the current mechanisms used to identify those activities are inefficient and therefore genuine customers who haven't done anything wrong get caught up in it. I presume both the service providers and the customers would like to get this need-gap addressed.
I'm currently developing a tool that will test website's bot protection levels - https://botmenot.com/
If you are concerned with bot detection, bot traffic, server response time, spam/fake content, web scraping, then check out this tool!
All the best on your development, Can you explain how you are addressing or planning to address the problem of false positives?
What are the things considered by the pattern matching systems of the fraud detection systems?
Are you talking about algorithm anomalies? Because I don't think this is a big issue. Imo of course before I get gutted.
But it happens all the time, Check the discussions about cloud service providers or on Twitter and it's not limited to cloud service as the other comment points out.