利用混沌特征分析的大型Web数据库异常检测

Abnormal Information Detection of Big Web Database Based on Chaotic Feature Analysis

为研究使用混沌分析的方法检测大型Web数据库的异常入侵特征新型问题,提出使用递归图分析的混沌特征分析方法检测Web数据库异常入侵。使用平均互信息算法和虚假最近邻点算法求取Web数据库信息流相空间重构的关键参数,使用递归图分析方法分析了各类异常入侵信号下真实Web数据库的检测。仿真结果表明平均互信息算法和虚假最近邻点算法能有效应用于对Web数据库信息流异常信号入侵检测的相空间重构中。递归图混沌分析的方法能有效检测出各类异常入侵特征,递归图中有规则图案,表明入侵信号和Web数据库信息流具有确定性成分存在,能对之实现有效检测和防御,研究结果证明检测算法能有效应用于网络数据安全检测实践。

Abnormal information detection of big Web database based on chaotic feature analysis was researched as a new problem. The recurrence plot analysis method was proposed in the application of detection of abnormal information for big Web database. The different types of abnormal information and signals were detected and analyzed in real Web database based on the recurrence plot analysis. Simulation result shows that the AMI and FNN algorithm can used as the calculation of the key parameters such as the delay time and embedded dimension for the phase space reconstruction in Web database flows application. Each type of intrusion abnormal signals can be detected with the recurrence plot analy-sis method. Results show that there are deterministic components in the Web database information flows, it can be de-tected and defended effectively by the proposed method, and research result shows the detection algorithm can be applied in the network database safety detection and defense in practice.