July 2014
·
4 Reads
Journal of Computational Information Systems
In order to discover importance order of trust relationships in decision of open networks, while adapting to network changes and avoiding subjectivities, an entropy gain based fusion model of decision attributes is proposed. Attribute entropy gain is calculated from trust decision samples. Attribute weights are designed in proportion to entropy gain. They are automatically adjusted when decision samples are refreshed. The fusion model follows the importance order of attributes in human decision, and adapt to network changes.