基于RTE和XGBoost组合模型的网络故障预测  

Network fault prediction based on RTE and XGBoost combined model

作　　者：杨龙 廖光忠[1,2] YANG Long;LIAO Guangzhong(School of Computer Science and Technology,Wuhan University of Science and Technology,Wuhan 430081,China;Hubei Province Key Laboratory of Intelligent Information Processing and Real-time Industrial System,Wuhan University of Science and Technology,Wuhan 430081,China)

机构地区：[1]武汉科技大学计算机科学与技术学院,湖北武汉430081 [2]武汉科技大学智能信息处理与实时工业系统湖北省重点实验室,湖北武汉430081

出　　处：《武汉大学学报（工学版）》2021年第8期755-760,共6页Engineering Journal of Wuhan University

基　　金：国家自然科学基金资助项目(编号:61502359)。

摘　　要：以某电信公司的网络设备告警日志作为研究基础,提出随机森林嵌入(random trees embedding,RTE)和极端梯度提升(extreme gradient boosting,XGBoost)的组合模型,并将其应用于对网络业务运行状态的故障预测。针对特征中噪声大的问题,采用RTE算法对告警日志数据进行特征转化,借助Bagging抽样方法减少噪声数据的影响;结合XGBoost算法的多分类功能建立预测模型。实验结果表明,相对于XGBoost算法、随机森林算法以及基于概率统计的贝叶斯算法,RTE+XGBoost组合模型的预测准确率分别提升了1.23%、3.44%、4.43%。Based on the network equipment alarm log of a telecommunications company, this paper proposed a combined model of RTE+XGBoost to apply to fault prediction of network business operation status. First of all, for noisy, the random trees embedding(RTE) algorithm was used to convert the warning log data to address the problem of high noise in the feature, and bagging sampling was used to reduce the impact of noise.Second, extreme gradient boosting(XGBoost) algorithm was exploited to establish multi-classification forecasting model. The experimental results show that the RTE+XGBoost combination model improves the prediction accuracy rates by 1.23%, 3.44% and 4.43%, respectively, compared to the XGBoost algorithm,random forest algorithm, and Bayesian algorithm based on probability statistics.

关 键 词：告警日志 RTE XGBoost 网络故障预测 

分 类 号：TP306.3[自动化与计算机技术—计算机系统结构]