Software to sort the friends from the spammers in social network services
July 25th, 2007

New Scientist reports on software that helps you determine whether you are being friended by a spammer in an online social network service:

Help is at hand for members of online communities who find themselves faced with suspicious requests from strangers to be accepted as a “friend”.

Social networking sites like MySpace and Facebook have spawned a new kind of nuisance. Spammers create realistic profile pages that make them look like ordinary users, and then persuade people to accept them as friends. This gives the spammers access to the victim’s inbox, which they flood with Viagra ads and pornography links.

The problem could overwhelm social networking sites. While there are filters that can distinguish spam emails from legitimate mail, it is hard to distinguish between spammers and legitimate users on a social networking site. “We can’t rely on any existing schemes for this environment,” says Aaron Zinman at the Massachusetts Institute of Technology.

To tackle the problem, Zinman, with his supervisor Judith Donath, created software that analyses MySpace pages for social behaviour by measuring how personalised a page is, how many shared photos and video clips it has, and the frequency of sent and received messages. Then the researchers looked at 800 MySpace pages picked at random, and graded them on how social – or, conversely, how promotional – they deemed the person to be. The software was then taught how to judge whether a profile is social or promotional by comparing the researchers’ data with measurements gleaned from its own analysis of the pages.

Fatal error: Call to undefined function sociable_html() in /home/permutype/ on line 36