A Multi-Perspective Fraud Detection Method For Multi- Participant E-Commerce Transactions

In our paper, we talk about how tricky it is to catch fraud in e-commerce when lots of people are involved. Usually, fraud detectors look at individual transactions, but they miss the bigger picture of how buyers, sellers, and middlemen interact. So, we came up with a new way to spot fraud in these complex transactions. Our method looks at transactions in different ways, like what's happening, how people are behaving, and who's connected to who. We use fancy computer programs and graphs to find subtle signs that something fishy might be going on. We also added a feature where we learn from past fraud cases to get better at catching new ones. We tested our method on real e-commerce data, and it did a good job at finding fraud in all sorts of situations.