..Dis-AE-LSTM: Generative Adversarial Networks for Anomaly Detection of Time Series Data
Sheng Mao, Jiansheng Guo, Taoyong Gu, Zhong Ma · 2020
Anomalies in time series data contain valuable information for engineers, scientists and entrepreneurs. However, the anomaly detection in time series is facing with the complexity caused by the large volume and high dimensions of data. To mitigate the problem, we propose an unsupervised anomaly detection method for time series data based on generative adversarial networks (GANs) with the long-short term memory recurrent neural networks (LSTM-RNN) as base modules. The nature of generative models is used to capture the distributions of high dimensional data, where the LSTM module is used to learn the temporal relationships. The proposed framework is designed with the capability to reconstruct input data after training, and the reconstruction errors based anomaly scores are assigned to each time step in the time series, which determines the anomalies. Experimental results show the superiorities and efficiency of the proposed method.