基于无监督学习的计算机通信网络异常数据流辨识方法

Method for abnormal data flow identification of computer communication network based on unsupervised learning

当前,计算机通信网络异常数据流辨识矩阵的辨识效率有限,导致漏辨个数不断增加。为此,文章对基于无监督学习的计算机通信网络异常数据流辨识方法进行了设计和验证。首先,提取了异常数据流特征,在多状态的背景下提升了辨识效率,同时设置了多状态特征辨识矩阵。其次,设计了无监督学习通信网络异常数据流辨识模型,采用了动态跟踪核验实现数据流辨识处理。测试结果表明,导入4组虚拟的辅助网络异常数据流指令在3个周期内实现了异常数据流辨识,最终将漏辨个数控制在6个以内,说明该方法在无监督学习的辅助和支持下更为精准与高效,应用效果得到了明显提升,且针对性较强,具有创新意义。

Currently,the identification efficiency of the abnormal data flow identification matrix in computer communication networks is limited,resulting in an increasing number of missed detections.Therefore,this article designs and verifies an unsupervised learning based method for identifying abnormal data flows in computer communication networks.Firstly,the abnormal data flow features were extracted to improve identification efficiency in a multi state background,and a multi state feature identification matrix was set up.Secondly,an unsupervised learning communication network anomaly data flow identification model was designed,and dynamic tracking verification was used to achieve data flow identification processing.The test results show that importing 4 sets of virtual auxiliary network abnormal data flow instructions achieved abnormal data flow identification within 3 cycles,and ultimately controlled the number of missed detections to within 6.This indicates that the method is more accurate and efficient with the assistance and support of unsupervised learning,and the application effect has been significantly improved,with strong targeting and innovative significance.