大数据驱动下网络入侵信号提取检测仿真

Simulation of Network Intrusion Detection and Extraction Based on Large Data Driven

大数据驱动下对网络入侵信号的提取检测,能够充分的保障大数据驱动下的网络安全。对网络入侵信号的提取检测,需要获取网络入侵提取的属性核,对网络入侵提取结果数据进行分类,完成入侵检测。传统方法定义网络入侵提取结果选取状态的分类的熵,给出各个熵的信息增益,但忽略了对网络入侵提取结果数据的分类,导致提取检测精度偏低。提出基于粗糙集-决策树结合的大数据驱动下的网络入侵信号提取检测模型。模型先利用粗糙集对大数据驱动下的网络中各个提取数据集中属性对应的取值进行离散化,获取网络入侵提取的属性核,利用决策树对新的网络入侵提取结果数据进行分类,初始化HMM模型的参数,将网络入侵信号提取检测的特征向量输入HMM,组建大数据驱动下的网络入侵信号提取检测模型。实验结果表明,所提模型建模精度较高,为保障大数据驱动下的网络安全奠定了基础。

This paper proposes a model for extraction detection of network instruction signal driven by large data based on integration of rough set and decision-making tree. Firstly, the discretization for corresponding value of con- centration property of each extraction data in network driven by large data utilizing rough set is carried out, then prop- erty nucleus of network instruction extraction are acquired. Moreover, the new data of extraction result are classified by utilizing decision-making tree, and the parameter of HMM model is initialized. The feature vector of the extraction detection is input into HMM. Finally, the model of extraction detection is built. Simulation results show that the mod- el has high accuracy.