云计算下的威胁数据挖掘模型仿真  

The Threat of The Cloud Data Mining Model Simulation

作　　者：刘宁[1] 管涛[1] 

机构地区：[1]郑州航空工业管理学院计算机科学与应用系,河南郑州450046

出　　处：《控制工程》2014年第6期958-961,965,共5页Control Engineering of China

基　　金：河南省科技厅基础与前沿技术研究计划项目(132300410186);河南省教育厅科技攻关项目(14B520067);河南省教育厅基础研究项目(2014B520067)

摘　　要：随着网络高级持续威胁为主的犯罪活动逐渐增加,使得云计算环境下的高级持续威胁数据检测,具有高纬度、非线性等特征,导致传统基于PSO辨识树的高级持续威胁数据挖掘过程中,采用的数据主元特征以及关联特征存在显著的波动,无法获取准确的数据挖掘结果。提出了一种基于改进流形学习算法的云计算下高级持续威胁数据挖据模型,使用非线性流形学习算法,降低云计算下高级持续威胁数据向量特征的维数,通过特征提取模块对高级持续威胁数据进行预处理,采用改进经典流形学习算法,加大样本散布密集区域高级持续威胁数据间的距离,缩短样本散布稀疏区域样本间的距离,促使云计算下高级持续威胁数据样本库的整体分布均匀化,实现云计算环境下高级持续威胁数据的准确挖掘。实验结果说明,所提方法能够准确挖掘出云计算环境下的高级持续威胁数据,具有较高的挖掘效率和精度。Continued threat as senior network crime increase gradually, make cloud computing environment, a senior continued threat data detection, has such characters as high latitudeS, nonlinear, the traditional tree based on PSO identification of senior continued threat data mining process, the principal characteristics and correlation characteristics exist significant fluctuations, unable to get accu- rate data mining results. Presents a cloud computing based on manifold learning algorithm advanced the constant threat data digging, ac- cording to the model using the nonlinear manifold learning algorithm, reduce the cloud top continued threat data vector feature dimen- sion, through the feature extraction module for senior continued threat data preprocessing, the modified classical manifold learning algo- rithm, enlarging the sample spread densely populated area senior continued threat the distance between the data, shortened the distance between sample spread sparse area, prompting the cloud top continued threat data sample library of whole distribution uniformity, realize the cloud computing environment, a senior continued threat data accurate mining. The experimental results indicate that the proposed method can accurately senior continued threat data mining a cloud computing environment, has the high mining efficiency and precision.

关 键 词：云计算 高级持续 威胁数据 挖掘模型 

分 类 号：F127[经济管理—世界经济]