基于LSTM的云环境异常智能检测方法研究

崔峰; 郭刚; 饶伟; 王敏红

基于LSTM的云环境异常智能检测方法研究

中铁信弘远(北京)软件科技有限责任公司, 北京 100089

基金项目:

中国铁路信息科技有限责任公司科技研究开发计划课题(2018B06)

详细信息

作者简介:
崔峰,助理工程师;郭刚,助理工程师。

中图分类号: U29;TP393
计量
- 文章访问数: 137
- HTML全文浏览量: 3
- PDF下载量: 22
出版历程
- 收稿日期: 2019-10-24

Anomaly intelligent detection method in cloud environment based on LSTM

Sinorail Hongyuan(Beijing) Software Technology Corporation Limited, Beijing 100089, China

摘要

摘要: 为提升运维人员响应速度,降低云环境异常对云上应用的影响,研究了一种基于长短期记忆(LSTM,Long Short-Term Memory)的云环境异常智能检测方法。通过将传统时间序列分析算法同LSTM神经网络相结合,实现在线预测云环境监控数据,并通过正态分布和贝叶斯推理定义预测波动范围,快速准确地判别云环境异常。在铁路云数据中心环境中进行测试验证,同其他时序预测方法的比较证明,本方法具有判别准确性高、对各种场景适用性强的优点,可为铁路大规模云数据中心智能运维实施提供一种有效的异常检测手段。
- 铁路云 /
- 异常检测 /
- LSTM神经网络 /
- 云环境
Abstract: In order to improve the response speed of operation and maintenance personnel and reduce the impact of cloud environment exceptions on cloud applications, this paper studied an anomaly intelligent detection method in cloud environment based on LSTM (Long Short-Term Memory).The paper combined traditional time series analysis algorithm with LSTM neural network to implement on-line prediction of cloud environment monitoring data, and defined the prediction fluctuation range by normal distribution and Bayesian inference to quickly and accurately identify cloud environment anomalies. Tested and verified in the railway cloud data center environment, Compared with other time series prediction methods, this method has the advantages of high discrimination accuracy and strong applicability to various scenarios. It can provide an effective anomaly detection method for the implementation of intelligent operation and maintenance of large-scale railway cloud data center.
- railway cloud /
- anomaly detection /
- neural network /
- cloud environment

HTML全文

参考文献(8)

[1]	高明星,刘洋,王晓莉,等. 铁路IaaS平台解决方案研究[J]. 铁路计算机应用, 2016, 25(12):15-19.
[2]	刘宇,王晓莉,高明星,等. 铁路信息系统自动化运维平台方案研究[J]. 铁路计算机应用, 2019, 28(1):34-37.
[3]	王鑫,吴际,刘超,等. 基于LSTM循环神经网络的故障时间序列预测[J]. 北京航空航天大学学报, 2018(4):772-784.
[4]	任智慧,徐浩煜,封松林,等. 基于LSTM网络的序列标注中文分词法[J]. 计算机应用研究, 2017, 34(5):1321-1324 , 1341.
[5]	Alex Graves, Jürgen Schmidhuber. Framewise phoneme classification with bidirectional LSTM and other neural network architectures[J]. Neural Netw, 2005, 18(5):602-610.
[6]	Shi X, Chen Z, Hao W, et al. Convolutional LSTM Network:a machine learning approach for precipitation nowcasting[C]//International Conference on Neural Information Processing Systems. 2015.
[7]	徐尧强,方乐恒,赵冬华,等. 基于LSTM神经网络的用电量预测[J]. 电力大数据, 2017(8):25-29, 14.
[8]	孙瑞奇. 基于LSTM神经网络的美股股指价格趋势预测模型的研究[D]. 北京:首都经济贸易大学, 2015.