Proud of Anat Haim, who successfully defending her Master’s project, “Classifying privacy behavior patterns in online social networks”, advised by Hadas Schwartz-Hassidim and Eran Toch. The project used supervised and unsupervised non-linear classification methods to analyze how users manage their exposure to other people in social networks.

For example, the diagram below shows one result of her study. The diagram portrays a decision tree (J48) that classify users according to their Agility – a measure for the diversity of audiences the users share their information with. We can see that the two most deciding factors are OSN literacy (the technical abilities that users possess to control their privacy) and bonding social capital (the weight that they give to bind with strong social ties).

Agility tree