Indirect weighted association rules mining for academic network collaboration recommendations
AusDM '12 Proceedings of the Tenth Australasian Data Mining Conference - Volume 134
| Hi-index | 0.00 |
SNSs pose numerous privacy issues that are reasonably well known and understood. Many issues boil down to the same problem: information makes it to the wrong audience. This problem is inherent to the design and business model of the current social network sites. Facebook is a case in point here. Its architecture is biased towards information disclosure to large audiences. Recently, Google has (re)entered the SNS arena with Google+ which seems to mitigate some of Facebook's 'flaws' by introducing the notion of 'circles' to designate audiences. Circles may improve the control users have over the information they share with their friends, family, colleagues, peer, etc, but is this a reason to safely flock to G+? Not quite. Both Facebook and Google not only aim to play a prominent role in the social network landscape, but also aim to become key players in the identity management landscape. They are inducing service providers to make use of their identity services (Facebook Connect, Google account) potentially offering access to the rich user profiles they have. The message that audience segregation matters and that individuals have multiple partial identities is thereby ignored which undermines privacy. In this lecture I will argue that privacy and sociality can be combined and that the architecture of Social Network Sites plays a role in this respect. I will further argue that the current business models of SNSs and their ambition to become digital identity hubs present serious privacy challenges. I will show an alternative, the privacy preserving social network site (Clique) developed within the EU FP7 project PrimeLife.