Bounded-Abstention Multi-horizon Time-series Forecasting

2026-02-04

AI Summary

This work addresses the lack of effective abstention mechanisms for highly uncertain predictions and the neglect of structural dependencies among multi-step outputs in time series forecasting. It formally introduces the problem of abstention learning in the multi-step prediction setting and proposes three abstention strategies tailored to the inherent structure of multi-step forecasts, along with their theoretically optimal policies. Building on this foundation, the authors design a dynamic abstention algorithm that adaptively decides, at each forecasting step, whether to abstain by leveraging inter-step correlations. Extensive experiments across 24 real-world datasets demonstrate that the proposed method significantly outperforms existing baselines, substantially enhancing prediction reliability and model trustworthiness.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Reasoning under Uncertainty: Relational Probabilistic Models
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods

Abstract

Multi-horizon time-series forecasting involves simultaneously making predictions for a consecutive sequence of subsequent time steps. This task arises in many application domains, such as healthcare and finance, where mispredictions can have a high cost and reduce trust. The learning with abstention framework tackles these problems by allowing a model to abstain from offering a prediction when it is at an elevated risk of making a misprediction. Unfortunately, existing abstention strategies are ill-suited for the multi-horizon setting: they target problems where a model offers a single prediction for each instance. Hence, they ignore the structured and correlated nature of the predictions offered by a multi-horizon forecaster. We formalize the problem of learning with abstention for multi-horizon forecasting setting and show that its structured nature admits a richer set of abstention problems. Concretely, we propose three natural notions of how a model could abstain for multi-horizon forecasting. We theoretically analyze each problem to derive the optimal abstention strategy and propose an algorithm that implements it. Extensive evaluation on 24 datasets shows that our proposed algorithms significantly outperforms existing baselines.

Problem

• multi-horizon forecasting
• learning with abstention
• time-series prediction
• structured prediction
• abstention strategy

Innovation

• multi-horizon forecasting
• learning with abstention
• structured prediction
• time-series forecasting
• abstention strategy