A Dynamic Trust Estimation Method for 'Persona' from the Human Relationship of Social Web

In this paper, we propose a trust estimation method for a parson by the rating of his or her active audience of social web such as Twitter. We call this method 'trust by the rating of a persona's active audience.' A 'persona', such as 'Display Name' of Twitter (with additional authentication) or 'Claimed ID' of OpenID, denotes an artificial identity to represent one's social position in the user-centric authorization. We utilize the 'persona' to be an object to accumulate the rating of trust for IT services for our trust estimation method. The most significant difference of our trust model and traditional one like PKI is that the trustworthiness of persona is not stable. Therefore, we focused on the implement methods for the sensitivity of the temporal change of a persona's trustworthiness. The basic idea of our 'trust by the rating of a persona's active audience' is similar to the Google Page Rank method. We also try to show that this kind of social web based trust has a possibility to be a global infrastructure for trust.