Disambiguating Identity through Social Circles and Social Data

This paper presents an approach to disambiguate extracted identity information relating to different individuals through the use of social circles. Social circles are generated through the extraction and pruning of social networks using the analysis of existing social data. Social data encompasses information such as images, videos and blogs shared within a social network. Identity information is extracted by involving the user in both selecting their key identity features for disambiguation, and validating the retrieved information. Our approach provides a methodology to monitor existing identity information, applicable to addressing such issues as identity theft, online fraud and lateral surveillance.