Good ideas and conversation. No ads, no tracking. Login or Take a Tour!
The "secret sauce" of machine learning algorithms is almost always "shit that makes no sense, but makes it perform well on our testing set." That said, I can buy such an algorithm being good at making the predictions it makes. Give it as a project to a machine learning class and I bet most everyone would come up with something that works pretty well. I don't buy anyone doing it I'm-willing-to-let-it-advise-the-cops-to-kick-someone's-door-down well though.