An application of rough set on the network security model

A new network security model based on the rough set classifier is proposed in the paper. The degree of dissatisfaction (DoD) is defined as the probability that a node belongs to the malicious node set. The evaluation of the DoD about a node includes the direct evaluation (the local DoD) and the indirect evaluation (the recommended DoD). To improve the computation accuracy and the efficiency of the DoD we use the rough set classifier combined with Bayesian classifier. Rough set theory is used for rule induction from incomplete data sets. Based on historical transaction records we calculate the local DoD (LDoD) and based on feedback recommendation records we calculate the recommended DoD (RDoD). Trading nodes are classified according to the DoD. Compared with existing trust models, experiment results indicated that the proposed model can obtain the higher examination rate of malicious nodes and the higher transaction success rate.