Social Networks Against Spam
from the whatever-works dept
While the various social networking services continue to push forward, some researchers are using the same basic concept of linking people to their friends to see if social networks can help fight spam. The idea is that most people get their email from a limited social network of people, and that network can be worked out by looking at the “to:” “from:” and “cc:” portions of your email. The theory being that if I know two people, they’re much more likely to know each other as well, and the system can link them, and assume that if they email each other it’s unlikely to be spam. Spam messages, however, have no such linkage, and are more easily picked out from the crowd. Of course, it doesn’t sound like this works all that well. You don’t get any false positives, but it seems to only recognize a little over 50% of all spam. The folks behind this research say it’s still useful, if used in combination with other spam filtering techniques.
Comments on “Social Networks Against Spam”
I’m not sure I understand the difference between this and white/grey/black-listing..?
The difference here is that they’re looking at the overall connections via a centralized system. So, it’s not just who you know, but the fact that if you know Alice and Bob, the system will assume it’s more likely that Bob and Alice know each other…