摘要
研究计算机网络大规模高维数据流异常数据挖掘方法,有效挖掘计算机网络大规模高维数据流异常数据,提升计算机网络安全性。使用基于Python网络爬虫的数据采集技术采集计算机网络大规模高维数据流,经软件总线模型完成数据流清洗预处理,降低数据流规模与维度后,利用基于枢纽现象与加权离群分数的离群数据挖掘算法,经计算机网络数据流数据对象K近邻查询、K近邻数据对象离群分数和求解与加权、区分度阈值生成等操作,获取计算机网络数据流异常数据,并通过构建卷积神经网络异常数据类型识别模型,有效识别异常数据类型。实验结果表明:该方法可有效挖掘计算机网络大规模高维数据流数据中存在的异常数据,异常数据挖掘与识别准确性较高,可显著提升计算机网络安全性。
In this study,the data mining method for anomalous large-scale high-dimensional data streams of computer networks is used to effectively mine anomalous data and improve the security of computer networks.Use the data collection technology based on Python web crawler to collect large-scale high-dimensional data streams of computer network,complete the data stream cleaning pretreatment through the software bus model,reduce the size and dimension of the data stream,use the outlier data mining algorithm based on hub phenomenon and weighted outlier score,and use the data stream data object K nearest neighbor query,K nearest neighbor data object outlier score and solution and weighting differentiation threshold generation and other operations are used to obtain abnormal data from computer network data streams,and a convolutional neural network abnormal data type recognition model is constructed to effectively identify anomalous data types.The experimental results show that this method can effectively mine anomalous data in large-scale high-dimensional data streams of computer networks,and the accuracy of anomalous data mining and recognition is high,which can significantly improve the security of computer networks.
作者
郑湘辉
张雪冰
Zheng Xianghui;Zhang Xuebing(College of Artificial Intelligence,HeFei College of Finance&Economics,Hefei,Anhui 230601,China)
出处
《黑龙江工业学院学报（综合版）》
2023年第8期105-110,共6页
Journal of Heilongjiang University of Technology(Comprehensive Edition)
关键词
加权离群分数
计算机网络
大规模
高维数据流
异常数据挖掘
数据清洗
weighted outlier score
computer networks
large-scale
high-dimensional data streams
data mining for anomalous data
data cleaning
