🤖 AI Summary
This paper addresses the “performative” problem in time-series forecasting—where predictions themselves causally influence the underlying data-generating process, inducing distributional shift and degrading long-horizon accuracy. We formally define the performative time-series (PeTS) forecasting task for the first time. To tackle it, we propose the Feature Performative Shift (FPS) framework, which jointly models delayed causal responses, dynamically reweights features based on their performative sensitivity, and incorporates distributionally robust optimization—yielding provably improved generalization error under causal awareness. FPS is model-agnostic and integrates seamlessly with mainstream architectures including LSTM, TCN, and Informer. Empirical evaluation on COVID-19 transmission dynamics and urban traffic flow forecasting demonstrates consistent superiority over state-of-the-art methods, achieving an average 12.7% reduction in MAE. Moreover, FPS significantly enhances prediction stability and robustness to policy interventions.
📝 Abstract
Time-series forecasting is a critical challenge in various domains and has witnessed substantial progress in recent years. Many real-life scenarios, such as public health, economics, and social applications, involve feedback loops where predictions can influence the predicted outcome, subsequently altering the target variable's distribution. This phenomenon, known as performativity, introduces the potential for 'self-negating' or 'self-fulfilling' predictions. Despite extensive studies in classification problems across domains, performativity remains largely unexplored in the context of time-series forecasting from a machine-learning perspective. In this paper, we formalize performative time-series forecasting (PeTS), addressing the challenge of accurate predictions when performativity-induced distribution shifts are possible. We propose a novel approach, Feature Performative-Shifting (FPS), which leverages the concept of delayed response to anticipate distribution shifts and subsequently predicts targets accordingly. We provide theoretical insights suggesting that FPS can potentially lead to reduced generalization error. We conduct comprehensive experiments using multiple time-series models on COVID-19 and traffic forecasting tasks. The results demonstrate that FPS consistently outperforms conventional time-series forecasting methods, highlighting its efficacy in handling performativity-induced challenges.
