A fusion model of overall trust relationship from multiple attributes in trusted networks based on dynamic entropy gain

An entropy gain based overall trust degree (OTD) computation model is proposed for large scale trusted networks. To overcome subjectivity in trust decision, attributes´ entropies and their gains are adopted to set attribute weights in the OTD fusion model. Attribute weights are dynamically adjusted by refreshing sample set and attribute entropy gains, to adapt to variations of network entities. Under the proposed method, attribute weights maintain relative stable but keep responses to network uncertainties.