Tailored Forecasting from Short Time Series via Meta-learning

TL;DR

This paper introduces METAFORS, a meta-learning approach that leverages related long time series to enable accurate forecasting of systems with limited historical data, demonstrated on chaotic systems.

Contribution

We propose a novel meta-learning framework that generalizes across related systems to improve short time series forecasting without needing contextual labels.

Findings

METAFORS reliably predicts short-term dynamics.

It captures long-term statistics effectively.

Works even with substantially different systems.

Abstract

Machine learning models can effectively forecast dynamical systems from time-series data, but they typically require large amounts of past data, making forecasting particularly challenging for systems with limited history. To overcome this, we introduce Meta-learning for Tailored Forecasting using Related Time Series (METAFORS), which generalizes knowledge across systems to enable forecasting in data-limited scenarios. By learning from a library of models trained on longer time series from potentially related systems, METAFORS builds and initializes a model tailored to short time-series data from the system of interest. Using a reservoir computing implementation and testing on simulated chaotic systems, we demonstrate that METAFORS can reliably predict both short-term dynamics and long-term statistics without requiring contextual labels. We see this even when test and related systems exhibit substantially different behaviors, highlighting METAFORS' strengths in data-limited scenarios.