Trust-based mechanism to pass confidential health information through a social network

Abstract

Web based social networks have gained a lot of concentration nowadays, and a large number of participants are readily available while communicating and sharing information. This concept has grown rapidly over the past several years. Social Networking sites make relationships among people and other social entities while formulating online communities. The user connectivity is based on trust. To establish connections, a user of a social network must advertise themselves to other users. Though Online Social Networking provides convenient platforms for sharing and spreading data among large audience, challenges come up with the security applications. Sensitive private information could be compromised if an account is unauthorizedly accessed. The scope of this research project is to protect private confidential information against discovery and disclosure, by unauthorized users. In order to fulfil this achievement, a trust-based system has been built on top of a social environment, by means of accomplishing the advantages of trust relationships. By extracting corresponding trust-related parameters of social network data, a network structure has been implemented. Afterwards a Bayesian Belief Network (BBN) was constructed and the filter plugin named Trusted Network for Gephi software was being designed using the modelled BBN. The user is able to run this filter on a particular social network and obtain the trusted network. The system was evaluated by performing a personal recommendation survey. Trusted Network filter as well as the Most Untrusted Nodes filter plugin was utilized in the evaluation process to assess real Facebook data. Results show that the system has a high accuracy. The main limitation identified is that human factors could not be measured through such a computerized system, which is a key factor that affects Trust . But three other major factors; popularity of an individual within his network, importance of an individual to his network, and the influence he can do for his network have been considered while obtaining a better performance in the system.