摘要
数据质量波动的监测有助于发现网络性能瓶颈或性能故障,对网络系统的稳定运行有很大帮助。为了主动发现数据指标中的异常波动现象,本文首先提出了一种多元高斯分布监测模型,用来检测离群点数据。经过数据预处理、模型训练和输出异常与定位等模块,最终本文实现了监测全网数据并筛选输出异常数据及位置的系统算法,最终模型准确率达到90%以上。
The monitoring of data quality fluctuation is helpful to find network performance bottleneck or performance fault, and it is helpful to the stable operation of network system. In order to detect the abnormal fluctuation in the data index actively, a multivariate Gaussian distribution monitoring model is proposed in this paper to detect outliers. After data preprocessing, model training and outputting anomaly and positioning modules, the system algorithm of monitoring the whole network data and filtering outlier data and location is realized. The accuracy of the model is over 90%.
引文
[1]胡晓洁.正态分布及其扩展综述[J].数学学习与研究,2014(3):92-94.
[2]傅惠民.多元正态分布整体推断方法[J].航空动力学报,2005,20(6):7-11.
[3]Banerjee O,Ghaoui L E,d’Aspremont A.Model selection through sparse maximum likelihood estimation for multivariate gaussian or binary data[J]//Journal of Machine Learning Research,2008,9(Mar):485-516.
[4]Li G,Rai A,Lee H,et al.Operational Anomaly Detection in Flight Data Using a Multivariate Gaussian Mixture Model[C]//PHM Society Conference.2018,10(1).