Probing the Robustness of Time-series Forecasting Models with CounterfacTS

arXiv:2403.03508v1 Announce Type: new
Abstract: A common issue for machine learning models applied to time-series forecasting is the temporal evolution of the data distributions (i.e., concept drift). Because most of the training data does not reflect such changes, the models present poor performance on the new out-of-distribution scenarios and, therefore, the impact of such events cannot be reliably anticipated ahead of time. We present and publicly release CounterfacTS, a tool to probe the robustness of deep learning models in time-series …

abstract arxiv concept cs.lg data distribution drift evolution forecasting issue machine machine learning machine learning models performance robustness series temporal training training data type