A Model for Inferring Trust for an Unknown User on Social Networking Sites

Abstract

The advent of technology has given birth to new ways of socializing in the form of web based social networks. The increase in the
number of users being associated with these networks like Facebook, LinkedIn or Orkut has drawn attention of many researchers in the past few
years. The reason is not the increasing number, but the way users share their personal and sensitive information without being aware of its
potential to be misused. If a malicious user or a non-trusted person gets access to our social network, he might cause a lot of damage to various
aspects of our personal life. Hence, allowing only trusted people to social networks is an important matter of concern. This gives rise to the
concept of trust, what trust is and on what basis one can trust another person on a web based social network. Previous studies have already shown
how trust ratings given by others can influence the overall trust rating. A lot of work has already been done in the field of inferring trust score or
trust value for users based on which another user can make decisions, but most of them have primarily focused on the trust ratings given to a
salable product for computing the trust score .This paper will be focusing on another model for inferring trust value between unknown members,
by first finding trusted paths between them and then using the trust ratings given by the immediate neighbors of the sink, that are also a part of
those trusted paths.

Keywords: trust rating; trust propagation; social networking sites; trust score; transitivity reduction;