互联网中混合入侵信息节点定位识别仿真

Localization and Identification of Hybrid Intrusion Information Nodes in Internet

对互联网中混合入侵信息节点的定位识别,能够有效抵御混合入侵节点的恶意攻击行为。对入侵信息节点的定位识别,需要通过类型相关节点属性,生成节点分类器,完成对入侵信息节点的定位。传统方法利用势函数计算混合入侵信息抽取的节点位置值,对各入侵节点的位置信任度进行综合,但忽略了生成节点分类器,导致节点定位识别精度偏低。在互联网环境下,提出基于多元分类的混合入侵信息节点识别方法。该方法首先在信誉阈值判定模型的基础上,引入第三方节点的间接可信度,将混合入侵信息中多种攻击类型相应的孤立点信任值进行整合,组建全部参与混合入侵信息节点的样本空间,生成节点分类器,在此基础上对未知类型的混合入侵信息节点进行分类。实验结果表明,所提方法能够准确识别节点定位,且节点定位识别效率较高。

A node recognition method for hybrid intrusion information based on multiple classification in Internet environment is proposed. The indirect reliability of third-party node based on decision model of reputation threshold value is introduced, then trust value of relevant isolated point of multiple attack types in hybrid intrusion information is integrated and carried out. In addition, the sample space of node for hybrid intrusion information with whole partic- ipation is built and the node classifier is generated. On that basis, the hybrid intrusion information nodes with un- known types are classified. Experimental results show that the method can focus recognition to node localization exact- ly. It has high efficiency for node localization and recognition.