Created a program to identify fake friends
Another useful program created for the analysis of social networks on the Internet. The development of the staff at the Sociable Media Group from the Massachusetts Institute of Technology is able to analyze user profiles in order to effectively identify spammers ( PDF ). Spammers are “fake” users who create accounts on a social network solely in order to become friends, send junk messages, promote some products. Now they will be calculated automatically.
To identify “fakes”, the program analyzes several parameters of a personal profile, including how much the page differs from standard templates, how many personal settings there are, how many personal photos and video clips are published, how often messages are sent and received.
During the test, the program analyzed 800 profiles on MySpace and made graphs of how each of them is “social” or, on the contrary, advertizing - as a percentage. This development can be integrated into a social network, so when you receive a request to enter the number of your friends from any user, you will immediately see this percentage.
Unfortunately, the developed program is unlikely to be a panacea. Independent experts warn that spammers are fully capable of pretending to be human beings and deceive this system. To effectively calculate the "parasites", you need not so much to analyze the content of their pages, as the structure of their social links. In theory, the circle of friends at the spammer should be different from the normal human.
To identify “fakes”, the program analyzes several parameters of a personal profile, including how much the page differs from standard templates, how many personal settings there are, how many personal photos and video clips are published, how often messages are sent and received.
During the test, the program analyzed 800 profiles on MySpace and made graphs of how each of them is “social” or, on the contrary, advertizing - as a percentage. This development can be integrated into a social network, so when you receive a request to enter the number of your friends from any user, you will immediately see this percentage.
Unfortunately, the developed program is unlikely to be a panacea. Independent experts warn that spammers are fully capable of pretending to be human beings and deceive this system. To effectively calculate the "parasites", you need not so much to analyze the content of their pages, as the structure of their social links. In theory, the circle of friends at the spammer should be different from the normal human.
')
Source: https://habr.com/ru/post/12605/
All Articles