Online Social Network (OSN) services have rapidly grown into a wide network and offer
users a variety of benefits. However, they also bring new threats and privacy issues to the
community. Unfortunately, there are attackers that attempt to expose OSN users’ private
information or conceal the information that the user desire to share with other users.
Therefore, in this research we develop a framework that can provide trusted data management
in OSN services. We first define the data types in OSN services and the states of
shared data with respect to Optimal, Under-shared, Over-shared, and Hybrid states. We
also identify the facilitating, detracting, and preventive parameters that are responsible for
the state transition of the data. In a reliable OSN service, we address that a user should be
able to set up his or her desired level of information sharing with a certain group of other
users. However, it is not always clear to the ordinary users how to determine how much
information they should reveal to others. In order to support such a decision, we propose
an approach for helping OSN users to determine their optimum levels of information
sharing, taking into consideration the payoffs (potential Reward or Cost) based on the
Markov decision process (MDP). As an extension of the MDP-based approach, we also
introduce a game theoretic approach, considering the interactions of OSN users and attackers
with conflicting interests whose decisions affect each other’s. Finally, after
developing the framework for the optimal data sharing on OSNs, we conduct several experiments
with attack simulation based on the proposed ideas and discuss the results. Our
proposed approach has the capability to allow a large amount of variables to be altered to
suit particular setups that an organization might have.