Annual Computer Security Applications Conference (ACSAC) 2012

Full Program »

Twitter Games: How Successful Spammers Pick Targets

Presentation
View File
pdf
886KB
Paper
View File
pdf
594KB

Vasumathi Sridharan
Indiana University - Bloomington
United States

Vaibhav Shankar
Indiana University - Bloomington
United States

Minaxi Gupta
Indiana University - Bloomington
United States

Abstract:
Online social networks (eg. Twitter) are an evolving plat-
form for communication on the Internet. Spam strategies
on Twitter have evolved signi cantly in this short period
of time. In this paper,we investigate strategies employed
by spammers on Twitter to reach relevant target audience.
Due to this targeted approach to spam, we also see unprece-
dented evidence of a large number of these spam accounts
forming relationships with other Twitter users, thereby be-
coming deeply embedded in the social network.
We analyze nearly 20 million tweets from about 7 million
Twitter accounts over a period of five days. We identify
a set of 14,230 spam accounts that managed to live longer
when compared to 73% of other spam accounts in our data
set. We characterize their behavior, type of tweets used,
how they target their audience and the applications that
are used. Signi cant portions of our fi ndings are contrary to
one of the previous work done by Thomas et al, suggesting
despite Twitter's e fforts to combat spam, spammers have evolved
in less than a year and most spammers exploit Twitter for
monetary bene fits. We identify four major strategies used
by spammers on Twitter that accounted for 84% of all spam
tweets in our dataset. One major strategy that accounted
for 56% of spam tweets (but only 15% of spam accounts)
involved the Amazon aliate program. Twitterfeed, a tool
that facilitates automated RSS feed, was found to be majorly
used by spammers for affiliate marketing.

 

Powered by OpenConf®
Copyright ©2002-2012 Zakon Group LLC