摘要
对异构网络信息中漂移数据流的检测,可有效提高网络信息数据的稳定性。对漂移数据流进行有效检测,需要先对属性分类划分,利用网络信息的不确定性解决数据流的漂移。传统方法利用随机决策树模型构建异构网络信息数据流集成分类器,但忽略了对数据流的属性分类,导致检测精度低。提出基于朴素贝叶斯理论的异构网络信息漂移数据流检测方法。首先估计异构网络信息未知数据流的不确定数值属性,结合自适应决策树节点的分割理论对不确定数值属性进行划分,将其转变为不确定分类属性,再结合朴素贝叶斯理论训练异构网络信息数据流基分类器,在合理处理网络数据流中不确定性的同时,有效解决异构网络信息数据流中隐含的偏移问题,完成对异构网络信息漂移数据流检测。实验结果表明,所提方法能够有效检测异构网络信息数据流漂移现象,且检测精度较高。
A detection method for drift data flow in heterogeneous network information based on naive Bayesian theory is proposed.Firstly,the uncertain value attribute of unknown data flow in the heterogeneous network information is estimated.The value attribute is divided integrated segmentation theory of decision tree node with self-adaption and converted into uncertain classification attribute.Then,the classifier is trained based on data flow integrated with the naive Bayesian theory.The uncertainty in network data flow is processed reasonably.Meanwhile,the drift problem implied in the data flow of heterogeneous network information is solved.Finally,the detection of drift data flow is completed.Experimental results show that the method can detect the data flow drift of heterogeneous network information effectively.It has high detection precision.
出处
《计算机仿真》
北大核心
2017年第3期357-360,共4页
Computer Simulation
关键词
异构网络信息
漂移数据流
数据流检测方法
Heterogeneous network information
Drift data flow
Data flow detection method
