Sensitive direct trust computation model for trusted networks
一种可信网络中敏感的直接信任度计算模型

Learning from features of human trust relationship such as forgetting, attenuation, relatively stable but dynamically sensitive, asymmetry and united space-time features of sensitivity, this paper proposed a direct trust computation model DTD based on a moving two-windowed mechanism. It adopted a wide stability window to obtain the stable part of DTD. It adopted a small sensitivity window to obtain the sensitive part of DTD. In the sensitivity window, it used two methods to describe the united space-time feature for variations of DTD to those of transaction satisfaction degree TSD. On the one hand, asymmetrically set TSDs negatively changed with high weights. On the other hand, dynamically and adaptively changed the window's width. Experiment results show that the proposed DTD model can maintain to be relatively stable but sensitive to minus variation of node behavior, and the sensitivity has a united space-time feature.