The Journal of Society for e-Business Studies (한국전자거래학회지)

Society for e-Business Studies (한국전자거래학회)
권호 표지
DOI QR코드

DOI QR Code

An Anomalous Sequence Detection Method Based on An Extended LSTM Autoencoder

확장된 LSTM 오토인코더 기반 이상 시퀀스 탐지 기법

  • Lee, Jooyeon (Division of Computer Science, Sookmyung Women's University) ;
  • Lee, Ki Yong (Division of Computer Science, Sookmyung Women's University)
  • Received : 2021.01.04
  • Accepted : 2021.02.19
  • Published : 2021.02.28
Download PDF
⟨ Previous Next ⟩

Abstract

Recently, sequence data containing time information, such as sensor measurement data and purchase history, has been generated in various applications. So far, many methods for finding sequences that are significantly different from other sequences among given sequences have been proposed. However, most of them have a limitation that they consider only the order of elements in the sequences. Therefore, in this paper, we propose a new anomalous sequence detection method that considers both the order of elements and the time interval between elements. The proposed method uses an extended LSTM autoencoder model, which has an additional layer that converts a sequence into a form that can help effectively learn both the order of elements and the time interval between elements. The proposed method learns the features of the given sequences with the extended LSTM autoencoder model, and then detects sequences that the model does not reconstruct well as anomalous sequences. Using experiments on synthetic data that contains both normal and anomalous sequences, we show that the proposed method achieves an accuracy close to 100% compared to the method that uses only the traditional LSTM autoencoder.

최근 센서 측정 데이터, 구매이력 등과 같이 시간 정보를 포함하는 시퀀스(sequence) 데이터가 다양한 응용에서 발생되고 있다. 주어진 시퀀스들 중 다른 시퀀스들과 매우 상이한 이상(anomalous) 시퀀스를 탐지하는 기법들은 지금까지 많이 연구되어왔으나 이들 대부분은 주로 시퀀스 내 원소들의 순서만을 고려하여 이상 시퀀스를 찾는다는 한계가 있다. 따라서 본 논문에서는 원소들의 순서와 원소들 간의 시간 간격 모두를 고려하는 새로운 이상 시퀀스 탐지 기법을 제안한다. 본 논문에서 제안하는 방법은 확장된 LSTM 오토인코더 모델을 사용한다. 이 모델은 시퀀스를 해당 시퀀스 내 원소들의 순서와 시간 간격 모두를 효과적으로 학습할 수 있는 형태로 변환하는 층을 추가로 가진다. 제안방법은 확장된 LSTM 오토인코더 모델로 주어진 시퀀스들의 특징을 학습한 뒤, 해당 모델이 잘 복원하지 못하는 시퀀스를 이상 시퀀스로 탐지한다. 본 논문에서는 정상 시퀀스와 이상 시퀀스를 혼합한 가상 데이터를 사용하여 제안 방법이 전통적인 LSTM 오토인코더만을 사용하는 방법과 비교하여 100%에 가까운 정확도를 나타냄을 보인다.

Keywords

References

  1. Boniol, P. and Palpanas, T., "Series2Graph: graph-based subsequence anomaly detection for time series," Proc. VLDB Endow, Vol. 13, No. 2, pp. 1821-1834, 2020. https://doi.org/10.14778/3407790.3407792
  2. Cai, S., Li, L., Li, Q., Li, S., Hao, S., and Sun, R., "UWFP-Outlier: an efficient frequent-pattern-based outlier detection method for uncertain weighted data streams," Applied Intelligence, Vol. 50, pp. 3451-3470, 2020.
  3. Cao, L., Yan, Y., Madden, S., Rundensteiner, E. A., and Gopalsamy, M., "Efficient discovery of sequence outlier patterns," Proc. VLDB Endow, pp. 920-932, 2019.
  4. Chalapathy, R. and Chawla, S., "Deep learning for anomaly detection: A survey," [online] Available: http://arxiv.org/abs/1901.03407, 2019.
  5. Chen, Z., Yeo, C. K., Lee, B. S., and Lau, C. T., "Autoencoder-based network anomaly detection," ireless Telecommunications Symposium, pp. 1-5, 2018.
  6. Chong, Y. S. and Tay, Y. H., "Abnormal Event Detection in Videos Using Spatiotemporal Autoencoder," Advances in Neural Networks-ISNN 2017, pp. 189-196, 2017.
  7. Ghrib, Z., Jaziri, R., and Romdhane, R., "Hybrid approach for Anomaly Detection in Time Series Data," 2020 International Joint Conference on Neural Networks (IJCNN), pp. 1-7, 2020.
  8. Hochreiter, S. and Schmidhuber, J., "Long short-term memory," Neural computation, Vol. 9, No. 8, pp. 1735-1780, 1997. https://doi.org/10.1162/neco.1997.9.8.1735
  9. Hu, J., Yang, B., Guo, C., and Jensen, C. S., "Risk-aware path selection with time-varying, uncertain travel costs: a time series approach," The VLDB Journal, Vol. 27, pp. 179-200, 2018. https://doi.org/10.1007/s00778-018-0494-9
  10. Kim, T.-Y. and S.-B. Cho., "Web traffic anomaly detection using C-LSTM neural networks," Expert Systems with Applications, Vol. 106, pp. 66-76, 2018. https://doi.org/10.1016/j.eswa.2018.04.004
  11. Lee, D. H. and Kim, K. H., "A LSTM Based Method for Photovoltaic Power Prediction in Peak Times Without Future Meteorological Information," The Journal of Society for e-Business Studies, Vol. 24, No. 4, pp. 119-133, 2019. https://doi.org/10.20428/jss.v24i4.1479
  12. Li, Z., Li, J., Wang, Y., and Wang, K., "A deep learning approach for anomaly detection based on SAE and LSTM in mechanical equipment," Int. J. Adv. Manuf. Technol., Vol. 103, pp. 499-510, 2019. https://doi.org/10.1007/s00170-019-03557-w
  13. Malhotra, P., Ramakrishnan, A., Anand, G., Vig, L., Agarwal, P., and Shroff, G., "LSTM-based encoder-decoder for multisensor anomaly detection," Proc. Anomaly Detection Workshop 33rd Int. Conf. Mach. Learn, 2016.
  14. Provotar, O. I., Linder, Y. M., and Veres, M. M., "Unsupervised Anomaly Detection in Time Series Using LSTM-Based Autoencoders," IEEE International Conference on Advanced Trends in Information Theory (ATIT), pp. 513-517, 2019.
  15. Wang, T., Duan, L., Dong, G., and Bao, Z., "Efficient Mining of Outlying Sequence Patterns for Analyzing Outlierness of Sequence Data," ACM Transactions on Knowledge Discovery, Vol. 14, No. 5, 2020.
  16. Wu, Z., Pan, S., Chen, F., Long, G., Zhang, C., and Yu, P. S., "A Comprehensive Survey on Graph Neural Networks," IEEE Transactions on Neural Networks and Learning Systems, 2020.
  17. Zhao, J., Li, Y., He, H., and Deng, F., "One-step Predictive Encoder-Gaussian Segment Model for Time Series Anomaly Detection," 2020 International Joint Conference on Neural Networks, pp. 1-7, 2020.