Anomaly detection of cloud virtual machine based on LSTM-GRBM

HU Yue; LIN Guo-yuan; Wei Guo-ying

Chinese

您当前的位置：

首页 >

文章列表页 >

Anomaly detection of cloud virtual machine based on LSTM-GRBM

更新时间：2026-09-04

- Anomaly detection of cloud virtual machine based on LSTM-GRBM
- Vol. 38, Issue 4, Pages: 46-51(2021)
- 作者机构：
  
  1. 中国矿业大学计算机科学与技术学院,江苏,徐州,221116
  2. 矿山数字化教育部工程研究中心机构,江苏,徐州,221116
  3. 南京大学新软件技术国家重点实验室,江苏,南京,210093
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2021
- 稿件说明：
移动端阅览
胡悦, 林果园, 蔚国莹. 基于LSTM-GRBM的云虚拟机异常检测[J]. 微电子学与计算机, 2021, 38(4): 46-51.

HU Yue, LIN Guo-yuan, Wei Guo-ying. Anomaly detection of cloud virtual machine based on LSTM-GRBM[J]. 2021, 38(4): 46-51.
胡悦, 林果园, 蔚国莹. 基于LSTM-GRBM的云虚拟机异常检测[J]. 微电子学与计算机, 2021, 38(4): 46-51. DOI：

HU Yue, LIN Guo-yuan, Wei Guo-ying. Anomaly detection of cloud virtual machine based on LSTM-GRBM[J]. 2021, 38(4): 46-51. DOI：

摘要

本文提出了一种对云虚拟机进行异常检测的模型LsGrbmAd.该模型首先通过长短期记忆网络(LSTM)捕获云虚拟机性能指标的时序特征，同时利用Dropout以防止数据过拟合；其次利用高斯玻尔兹曼机(GRBM)得出自由能；最后利用得出的自由能与训练阶段得到的参数基准模型进行对比，判断云虚拟机是否出现异常.实验表明该模型能够对云虚拟机进行异常检测，且准确率有较大的提升.

Abstract

A model of Long Short Term Memory and Gaussian Boltzmann Machine Anomaly Detection(LsGrbmAd) is proposed in this paper. Firstly

the model captures the timing characteristics of cloud virtual machine performance indicators through long short term memory LSTM)

and uses Dropout to prevent data overfitting. Secondly

the free energy is obtained by using The Gaussian Boltzmann machine (GRBM). Finally

the free energy obtained is compared with the parameter benchmark model obtained in the training stage to judge whether there are abnormalities in the cloud virtual machine. The experiment shows that the model can detect the abnormity of the cloud virtual machine

and the accuracy is improved greatly.

关键词

Keywords

references

MONTUFAR G, AY N. Refinements of universal approximation results for deep belief networks and restricted Boltzmann machines[J]. Neural Computation, 2011, 23(5): 1306-1319. DOI: 10.1162/NECO_a_00113.

权鹏宇. 基于上下文的虚拟机异常检测研究[D]. 昆明: 昆明理工大学, 2018. QUAN P Y. A research on virtual machine anomaly detection based on context[D]. Kunming: Kunming University of Science and Technology, 2018.

叶青, 黄强, 聂斌, 等. 一种自适应的高维离群点识别方法[J]. 《广西师范大学学报》(自然科学版), 2020, 38(2): 107-114. DOI: 10.16088/j.issn.1001-6600.2020.02.012. YE Q, HUANG Q, NIE B, et al. An adaptive high-dimensional outlier recognition method[J]. Journal of Guangxi Normal University (Natural Science Edition), 2020, 38(2): 107-114. DOI: 10.16088/j.issn.1001-6600.2020.02.012.

LE ROUX N, BENGIO Y. Representational power of restricted Boltzmann machines and deep belief networks[J]. Neural Computation, 2008, 20(6): 1631-1649. DOI: 10.1162/neco.2008.04-07-510.

张健, 丁世飞, 张楠, 等. 受限玻尔兹曼机研究综述[J]. 软件学报, 2019, 30(7): 2073-2090. DOI: 10.13328/j.cnki.jos.005840. ZHANG J, DING S F, ZHANG N, et al. Restricted Boltzmann machines: a review[J]. Journal of Software, 2019, 30(7): 2073-2090. DOI: 10.13328/j.cnki.jos.005840.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Microservices anomaly detection based on the correlation of monitoring metrics

A Network Traffic Anomaly Detection Method Based on Non-parametric Statistical Theory

Aerial small object detection algorithm based on improved RT-DETR

Monte carlo analysis of V-layout trench SiC neutron detector

Lightweight countermeasures against Spectre attacks on the branch prediction unit

Related Author

WANG Chenrui

JIANG Ying

YANG Tianqing

DING Fan

YANG Yue

LI Jun

Related Institution

Yunnan Provincial Key Laboratory of Artificial Intelligence, Kunming University of Science and Technology

School of Information Engineering and Automation, Kunming University of Science and Technology

School of Big Data, Baoshan University

Institute for Pattern Recognition & Artificial Intelligence, HUST

AI问答

Postal code：100079
Tel：（010）53879206 Email：tmw@bjxintong.com.cn
Technical support is provided by Beijing Founder electronics co., LTD 京ICP备09082226号-64 京公网安备11010602201714号
It is recommended to read the content of this site in Chrome&IE9+. Please switch to extreme mode in browser 360.
Cookies We use cookies to help provide and enhance our service and tailor content. By continuing, you agree to the use of cookies.

⁰