【摘要】 对物联网中的入侵信号进行预警优化在提高物联网安全性能方面具有重要意义。由于物联网所处环境及特有的网络特征,使得网络传输特征多样。传统的预警方法主要通过网络传输特征确定其入侵信号特征,在根据信号特征进行预警,未考虑特征属性对入侵信号检测结果的影响,导致结果不准确的问题。提出一种适用于物联网环境的改进贝叶斯网络入侵信号预警方法研究,利用相似度选择网络数据的特征属性,提取关键特征向量,降低冗余度及噪音,优化朴素贝叶斯理论对物联网络入侵信号预警方法的分类性能,有效地实现对物联网络数据入侵信号的预警。仿真结果表明,提出的方法能够有效地实现物联网络入侵信号预警,在精确度和误差率控制方面具有明显优势。更多还原

【Abstract】 Early-warning optimization of intrusion signals is significant in aspect of improving security performance of the internet of things. In this paper,we proposed an early warning method of intrusion signal applying to environment of the internet of things based on modified Bayesian network. Firstly,we used the similarity degree to select feature attribute of network data. Then we extracted the key feature vector and reduced the redundancy and noise.Moreover,we optimized the naive Bayes theory to classification performance of intrusion signal early warning method in the internet of things. Finally,we achieved the early warning of intrusion signal effectively. The simulation results show that the method mentioned above can achieve the early warning of intrusion signal effectively. It has obvious advantages in aspect of precision and error rate control.更多还原