Abstract: With the sharing of images on social media such as Facebook, twitter, etc. increases, maintain their privacy becomes the major problem. As user shares their private images on social sites, people expect more tools to allow them to regain control over their privacy. By considering this need, we propose an Adaptive Privacy Policy Prediction (A3P) system which provides user convenient privacy settings by automatically generating personalized policies. To define users’ privacy preferences we consider the different factors such as social environment, personal characteristics, image content and metadata. We also consider the side information of the user such as contact list, visitors view and previous settings. For the images being uploaded, we define the best available privacy policy for the user based on the users’ available history on the site. For that we propose a two level framework. A3P system relies on the image classification framework for image categories which may be associated with similar policies and on a policy prediction algorithm to automatically generate a policy for each newly uploaded image, also according to users’ social features. We increase the efficiency by using the side information of the user.

Keywords: Social media, content sharing sites, metadata, policy mining, policy prediction.