一种采用串行自编码器的时序数据异常检测方法
作者:
作者单位:

1.山东财经大学计算机科学与技术学院;2.山东省数字媒体技术重点实验室;3.山东省未来智能金融工程实验室;4.山东大学软件学院

作者简介:

通讯作者:

中图分类号:

TP391

基金项目:

国家自然科学基金项目(61873145,61802229);山东省自然科学省属高校优秀青年人才联合基金项目(ZR2017JL029);山东省高等学校青创科技支持计划(2019KJN045)


A serial AutoEncoders based method for detecting time series anomalies
Author:
Affiliation:

School of Computer Science and Technology, Shandong University of Finance and Economics

Fund Project:

the National Natural Science Foundation of China (61873145, 61802229), Natural Science Foundation of Shandong Province for Excel-lent Young Scholars (ZR2017JL029), Science and Technology Innovation Program for Distinguished Young Scholars of Shandong Province Higher Education Institutions (2019KJN045)

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    基于深度学习的时序数据异常检测模型大多采用循环神经网络或长短期记忆网络来捕捉时序依赖性,并利用自编码器重构数据, 进而实现时序数据的异常检测. 虽然此类检测模型实现了较高的异常检测率, 但它们的网络结构复杂, 导致模型的计算效率较低. 为提高模型的计算效率, 提出了一种基于串行自编码器的异常检测模型SAE-AD. 该模型仅包含两个结构简单的自编码器(AE1 和AE2), 其所含参数量较少, 且训练目标较为简单, 从而加快了模型的计算效率. 通过将自编码器AE1和AE2串行拼接, 即AE1的输出作为AE2的输入, 有效提高了AE2的解码器对正常数据特征的解码能力, 有助于提升模型的检测准确率. 实验结果表明, 相较于其他新近提出的异常检测模型, SAE-AD模型具有更高的精确率、召回率和F1值.

    Abstract:

    Aiming to detect time series anomalies, deep learning methods generally use Recurrent Neural Network or Long Short Term Memory to capture temporal dependency, and adopt AutoEncoder to reconstruct data. Although they work well for detecting anomalies, the network structures of these methods are complex, resulting in slow computational efficiency. In order to improve the computational efficiency, this paper proposes a method called Serial AutoEncoders based Anomaly Detection (SAE-AD) which contains two AutoEncoders (AE1 and AE2) with simple structure. Due to the simplicity, there are a few training parameters and its training objectiv function is relatively simple, which speeds up the computation. In addition, the output of AE1 is fed into AE2 to improve the decoding ability of the decoder of AE2. This way of serial training makes SAE-AD achieve better detection accuracy. Experiment results show that the proposed method has better precision, recall, F1 score than several state-of-art anomaly detection methods.

    参考文献
    相似文献
    引证文献
引用本文
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2022-03-01
  • 最后修改日期:2023-01-16
  • 录用日期:2022-07-06
  • 在线发布日期: 2022-07-30
  • 出版日期: