A Trust Rating Method for Information Providers over the Social Web Service: A Pragmatic Protocol for Trust among Information Explorers and Information Providers

In this paper, we propose a trust rating method for information providers using the structure of human relationships and meta-level communication protocol over the social web service. In our first approach, we proposed a trust rating method based on the number of followers on Twitter and the similarity of activity of active Twitter users. However, the results of our experimentation showed that this approach does not provide an effective trust rating system because shrewd software robots get very high trust ratings under this method. As a reflection of the first approach, our second approach is that a trust rating method requires more explicit and precise evidence of trust. We consider that a trust rating system should meet the following conditions. (1) The system provides some methods to prove the person who is trusted has the ability to perform the trusted task. (2) The system provides some motivation for the people to maintain trusted behavior. (3) The system provides explicit meta-level communications such as 'I will trust you', and 'All right' and the results cause a change of the state and reputation of the parties involved. As an example of this method, we put forward a trust rating system for a job-hunting system for IT-engineers. We also discuss how to avoid the phenomenon of information cascading, which is a typical problem of collective intelligence systems over the social web.