Harmful data enhanced anomaly detection for quasi-periodic multivariate time series

Multivariate quasiperiodic time series (MQTS) anomaly detection has demonstrated significant potential across variouspractical applications, including health monitoring, intelligent maintenance, and quantitative trading. Recent research hasintroduced diverse methods based on autoencoders (AEs) and generative adversarial networks (GANs) that learn latentrepresentations of normal data and subsequently detect anomalies through reconstruction errors. However, anomalous trainingset data can cause model pollution, which harms the ability to of the utilized model reconstruct normal data. The currentdata extreme imbalance creates an enormous challenge in terms of stripping out these anomalies. In this paper, we propose aGAN-based multivariate quasiperiodic time series anomaly detection method called IGANomaly (I represents isolation). Thismethod isolates normal and harmful samples via pseudolabeling and then learns harmful data patterns to enhance the processof reconstructing of normal samples. First, the reconstruction error and potential feature distribution are jointly analyzed.Bimodal dynamic alignment is achieved through multiview clustering, thus overcoming the limitation of unidimensionaldetermination.Second,dualreconstructionconstraintsforthegeneratorandagradientpenaltymechanismforthediscriminatorare constructed. While maintaining the reconstruction quality achieved for normal samples, the propagation path of abnormal features is actively perturbed through a gradient inversion strategy. On three public datasets, IGANomaly achievesF1scoresof 0.811, 0.846, and 0.619, demonstrating an average improvement of 18.9% over the best baseline methods.