Contrastive Time Series Forecasting with Anomalies
This addresses the challenge of distinguishing between forecast-relevant and irrelevant anomalies in time series data, which is incremental as it builds on existing forecasting methods with a novel regularization approach.
The paper tackled the problem of time series forecasting in the presence of anomalies, where standard models often misclassify short-lived noise or persistent shifts, and proposed Co-TSFA, a regularization framework that improved performance under anomalous conditions while maintaining accuracy on normal data across Traffic, Electricity, and cash-demand datasets.
Time series forecasting predicts future values from past data. In real-world settings, some anomalous events have lasting effects and influence the forecast, while others are short-lived and should be ignored. Standard forecasting models fail to make this distinction, often either overreacting to noise or missing persistent shifts. We propose Co-TSFA (Contrastive Time Series Forecasting with Anomalies), a regularization framework that learns when to ignore anomalies and when to respond. Co-TSFA generates input-only and input-output augmentations to model forecast-irrelevant and forecast-relevant anomalies, and introduces a latent-output alignment loss that ties representation changes to forecast changes. This encourages invariance to irrelevant perturbations while preserving sensitivity to meaningful distributional shifts. Experiments on the Traffic and Electricity benchmarks, as well as on a real-world cash-demand dataset, demonstrate that Co-TSFA improves performance under anomalous conditions while maintaining accuracy on normal data. An anonymized GitHub repository with the implementation of Co-TSFA is provided and will be made public upon acceptance.