Loading...
|
Please use this identifier to cite or link to this item:
https://nccur.lib.nccu.edu.tw/handle/140.119/159034
|
| Title: | 基於生成對抗網路之時間序列異常偵測
Time Series Anomaly Detection based on Generative Adversarial Network |
| Authors: | 林柏辰
Lin, Po-Chen |
| Contributors: | 周珮婷
張志浩
林柏辰
Lin, Po-Chen |
| Keywords: | 生成對抗網路
時間序列
異常偵測
深度學習
Generative Adversarial Network
Time Series
Anomaly Detection
Deep Learning |
| Date: | 2024 |
| Issue Date: | 2025-09-01 14:48:27 (UTC+8) |
| Abstract: | 隨著物聯網系統的快速發展，人們愈來愈依賴使用傳感器獲取各種時間序列資料以執行自動化任務，因此辨識其中的異常狀況成為一個重要議題。然而由於時間序列特有的時間依賴性，異常偵測是項充滿挑戰且複雜的任務。本研究以深度學習方式，提出一種基於生成對抗網路之單變量時間序列異常偵測模型架構 CLAWGANdiv。透過訓練完成之生成網路重構出正常子序列，並計算出測試集中滑動窗口子序列之異常分數，藉此判定其是否異常。此外，為了能有效地捕捉資料中的時間依賴性，模型採用雙向長短期記憶網路做為主要架構，並結合注意力機制與卷積層以捕捉資料特徵。為驗證所提出模型之有效性，本研究使用 Numenta Anomaly Benchmark 資料集與 Yahoo Webscope 資料集中的部分資料，進行後續參數調整與其他模型之比較，從中取得優異的結果並針對實驗結果進行討論。
With the development of Internet of Things (IoT) systems, humankind has become increasingly dependent on sensor devices to acquire various time series data for task automation, making anomaly detection an extremely important issue. However, due to the highly complex temporal correlations of time series data, detecting anomalies might be particularly challenging. This research proposes CLAWGANdiv, a univariate time series anomaly detection approach based on Generative Adversarial Networks (GANs). The model reconstructs normal time series through a trained Generator and calculates anomaly scores for subsequences extracted with a sliding window in the test set to determine whether they are anomalous or not. To capture the temporal correlations of time series, we use Bidirectional Long Short-Term Memory network (BiLSTM) as the main architecture, complemented by an attention mechanism and convolutional layers to capture data features. To validate the effectiveness of the proposed method, we adjusted parameters and compared results with other models using a partial Yahoo Webscope dataset and Numenta Anomaly Benchmark (NAB) dataset. The results show that our approach can effectively detect anomalies, and further discussion will be based on these experimental results. |
| Reference: | Arjovsky, M., Chintala, S., and Bottou, L. (2017). Wasserstein gan.
Bashar, M. A. and Nayak, R. (2020). Tanogan: Time series anomaly detection with generative adversarial networks. CoRR, abs/2008.09567.
Cuturi, M. and Blondel, M. (2018). Soft-dtw: a differentiable loss function for time-series.
Elman, J. L. (1990). Finding structure in time. Cognitive science, 14(2):179–211.
Endres, D. and Schindelin, J. (2003). A new metric for probability distributions. IEEE Transactions on Information Theory, 49(7):1858–1860.
Geiger, A., Liu, D., Alnegheimish, S., Cuesta-Infante, A., and Veeramachaneni, K. (2020). Tadgan: Time series anomaly detection using generative adversarial networks.
Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., and Bengio, Y. (2014). Generative adversarial nets. Advances in neural information processing systems, 27.
Gulrajani, I., Ahmed, F., Arjovsky, M., Dumoulin, V., and Courville, A. (2017). Improved training of wasserstein gans.
Heusel, M., Ramsauer, H., Unterthiner, T., Nessler, B., and Hochreiter, S. (2017). Gans trained by a two time-scale update rule converge to a local nash equilibrium. Advances in neural information processing systems, 30.
Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory. Neural computation, 9(8):1735–1780.
Kullback, S. and Leibler, R. A. (1951). On information and sufficiency. The Annals of Mathematical Statistics, 22(1):79–86.
Laptev, N. and Amizadeh, S. (2019). A labeled anomaly detection dataset s5 yahoo research, v1. https://webscope.sandbox.yahoo.com/catalog.php?datatype=s&did=70.
Lavin, A. and Ahmad, S. (2017). The numenta anomaly benchmark (white paper). The Numenta Anomaly Benchmark [White paper].
LeCun, Y., Bottou, L., Bengio, Y., and Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11):2278–2324.
Li, D., Chen, D., Shi, L., Jin, B., Goh, J., and Ng, S.-K. (2019). Mad-gan: Multivariate anomaly detection for time series data with generative adversarial networks.
Mallasto, A., Montúfar, G., and Gerolin, A. (2019). How well do wgans estimate the Wasserstein metric?
Mirza, M. and Osindero, S. (2014). Conditional generative adversarial nets. arXiv preprint arXiv:1411.1784.
Miyato, T., Kataoka, T., Koyama, M., and Yoshida, Y. (2018). Spectral normalization for generative adversarial networks.
Radford, A., Metz, L., and Chintala, S. (2015). Unsupervised representation learning with deep convolutional generative adversarial networks. arXiv preprint arXiv:1511.06434.
Salimans, T., Goodfellow, I., Zaremba, W., Cheung, V., Radford, A., Chen, X., and Chen, X. (2016). Improved techniques for training gans. In Lee, D., Sugiyama, M., Luxburg, U., Guyon, I., and Garnett, R., editors, Advances in Neural Information Processing Systems, volume 29. Curran Associates, Inc.
Schlegl, T., Seeböck, P., Waldstein, S. M., Schmidt-Erfurth, U., and Langs, G. (2017). Unsupervised anomaly detection with generative adversarial networks to guide marker discovery. CoRR, abs/1703.05921.
Stanczuk, J., Etmann, C., Kreusser, L. M., and Schönlieb, C.-B. (2021). Wasserstein gans work because they fail (to approximate the wasserstein distance).
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., and Polosukhin, I. (2023). Attention is all you need.
Wu, J., Huang, Z., Thoma, J., Acharya, D., and Gool, L. V. (2018). Wasserstein divergence for gans.
Zhu, J.-Y., Park, T., Isola, P., and Efros, A. A. (2017). Unpaired image-to-image translation using cycleconsistent adversarial networks. Proceedings of the IEEE international conference on computer
vision. |
| Description: | 碩士
國立政治大學
統計學系
111354012 |
| Source URI: | http://thesis.lib.nccu.edu.tw/record/#G0111354012 |
| Data Type: | thesis |
| Appears in Collections: | [統計學系] 學位論文
|
Files in This Item:
| File |
Description |
Size | Format | |
|---|
| 401201.pdf | | 5129Kb | Adobe PDF | 0 | View/Open |
|
All items in 政大典藏 are protected by copyright, with all rights reserved.
|
