Abstract: In this paper keep the information of social graphs (SGs) available to their related business services. Some users also hide or limit the information of their connections from the public in social media platforms due to privacy concerns. Accessing these SGs is getting more difficult and costly in today’s online social networks, and novel applications using SGs become almost impossible to be offered independently by third-party practitioners and individuals. However, billions of user shared images are generated by the folks in many social networks daily, and this particular form of user data is indeed very accessible to others due to the nature of online image sharing. In addition, the paper show that privacy recommends models can be further enhanced by utilizing user enhanced privacy sentiment for mass customization. In this thesis detecting user improved privacy approach and privacy managing models can be automatically tailored specific to the privacy sentiment and needs of the user.

Keywords: Big data, Connection Discovery, Social network Service, User-shared images.