Syntactic analysis for monitoring personal information leakage on social network services: A case study on Twitter

Abstract

Social network services such as Twitter and Facebook can be considered as a new media different from the typical media group. The information on social media spread much faster than any other traditional news media due to the fact that people can upload information with no constrain to time or location. Because of this reason, people got fascinated by SNS and it sinks into our life. People express their emotional status to let others know what they feel about information or events. However, there is a high possibility that people not only share information with others, but also they expose personal information unintentionally such as place to live, phone number, date of birth, and more. This will be serious problem if someone has impure mind. It is actually happening in cyber-stalking, offline stalking or others. There are also many spam messages in SNS because of the fact that information in SNS spread much faster than any other media and it is easy to send a message to others. In other words, SNS provides vast backbone environment to spammers to hunt normal pure users. In order to prevent information leakage and detect spam messages, many researchers traditionally have been studied for monitoring email systems, web blogs, and so on. In this paper, we dealt with text message data in Twitter which is one of the most popular social network services over the world in order to reveal various hidden patterns. Twitter data is severely dangerous to organizations and more is that anyone who has Twitter account can access to any users by "following" function. The following function does not require permission from the requested person to confirm to ready their timelines. This study will be focused on the user to whom exchange text messages and what types of information they reciprocated with others by monitoring 50 million tweets on November in 2009 which was collected by Stanford University. © 2013 Springer-Verlag.