Abstract:
The number of people using the Internet as a main source of information is constantly growing.
And since the content of some of this information could be fraudulent or unreliable, the need for a method to determine its trustworthiness is essential. This research investigates the factors that contribute to trustworthiness of information in web pages, with each factor converted to an agent called factor analysis agent that generates a representative value for each factor. We have classified these factors into three bands according to human perception priorities. The research generates a computational trust model based on factors classification as a core of the Content Trust Agent (CTA). The CTA calculates a trust value of information in a given webpage depending on the factor analysis agents’ values that communicate between each other and CTA agent to produce our Multi-Agent System (MAS). This work evaluates the outputs of the trust model of our MAS that generate a trust value in a percentage and shows how much the trust value of our model is closed to human perception (from expert users).
Description:
CD, no of pages 133, 26597, informatics 3/2013