基于贵州联通宽带数据的网络异常检测与预测模型研究  

Research on Network Anomaly Detection and Prediction Model Based on Guizhou Unicom Broadband Data

作　　者：王谦 WANG Qian(China United Network Communications Co.,Ltd.,Guizhou Branch,Guiyang 550000,China)

机构地区：[1]中国联合网络通信有限公司贵州分公司,贵阳550000

出　　处：《移动信息》2024年第12期4-6,10,共4页Mobile Information

摘　　要：针对贵州联通宽带网络中的异常情况,文中提出了一种基于宽带数据的网络异常检测与预测模型。首先,收集并处理贵州联通的宽带数据,通过数据清洗、特征提取及异常数据标注,确保数据的有效性和准确性。其次,设计网络异常检测模型,利用特征选择和超参数调优,提升模型的检测精度和鲁棒性。最后,构建网络异常预测模型,结合时间序列分析技术,预测未来可能出现的网络异常现象。实验结果表明,该模型具有较高的准确性和稳定性,为提升宽带网络的可靠性提供了技术支持。Aiming at the anomaly situation in Guizhou Unicom̓s broadband network,a network anomaly detection and prediction model based on broadband data is proposed in this paper.First,the broadband data of Guizhou Unicom is collected and processed,and the validity and accuracy of the data are ensured through data cleaning,feature extraction and anomaly data annotation.Secondly,the network anomaly detection model is designed,and the detection accuracy and robustness of the model are improved by using feature selection and hyperparameter tuning.Finally,the network anomaly prediction model is constructed,and the possible network anomaly phenomena in the future are predicted by combining time series analysis technology.The experimental results show that the model has high accuracy and stability,providing technical support for improving the reliability of broadband networks.

关 键 词：宽带数据 网络异常 异常检测 预测模型 

分 类 号：TN913.2[电子电信—通信与信息系统]