Nowadays there are various Internet applications, such as blog and RSS blossoming on the Internet, among which the online social network service is the most popular and fast growing. Sina microblog is one of the widely used social network platform in China. However, with its prosperity and open nature the so-called “zombie fans problem” also come out: lager amount of accounts (known as socialbots) were registered by automated programs, and were used by some organizations such as public relation (PR) company to make profits by propagating biased or malicious information, which annoyed the common human users badly.

This work investigates into collective detection approach to discriminate the socialbots from normal human users. Features, such as tweet interval of time (TIT), weigh...[ Read more ]

Nowadays there are various Internet applications, such as blog and RSS blossoming on the Internet, among which the online social network service is the most popular and fast growing. Sina microblog is one of the widely used social network platform in China. However, with its prosperity and open nature the so-called “zombie fans problem” also come out: lager amount of accounts (known as socialbots) were registered by automated programs, and were used by some organizations such as public relation (PR) company to make profits by propagating biased or malicious information, which annoyed the common human users badly.

This work investigates into collective detection approach to discriminate the socialbots from normal human users. Features, such as tweet interval of time (TIT), weight of users, integrity degree are used. Our strategy succeeds at detecting most of the socialbots while only a small percentage of normal users are misclassified.

Keywords: socialbots, microblog