Sliding-Window Signatures as Features for Time Series Forecasting
Abstract
The evolution of a time series is often influenced not only by its past values but also by external covariates, giving rise to forecasting with exogenous variables. Nonlinear and delayed effects between covariates and the target series make this task particularly challenging. We propose a forecasting framework based on Ridge regression with sliding-window signature features, combining the computational efficiency of (semi-)parametric models with the flexibility of nonparametric approaches. These features capture complex dependencies between the target and the exogenous variables without relying on prior assumptions about the form of these relationships or the influence of past observations. We establish theoretical guarantees in the discrete-time setting, including universality of approximation and stationarity for the sliding-window transform. To ease implementation, we introduce a novel sequential algorithm that exploits window overlap for efficient signature updates. The method is evaluated on both synthetic and real electricity demand data. Results show that signature features effectively encode temporal and nonlinear dependencies, yielding accurate forecasts competitive with those based on expert knowledge and recent neural network approaches.
Related
- Encoding Time Series Dynamics with Sliding-Window Signatures for Online Forecasting
- Moving beyond stationarity: A sliding window z-score criterion for response time outlier exclusion
- Poincaré's Inequality on Sliding-Window Occupation Count Kernels and the Optimal Normalized Spectral-Gap Constant
- Informed forecasting: Leveraging auxiliary knowledge to boost large language model performance on time series forecasting
- Hierarchical Time Series Forecasting with Operator-Valued Kernels
BibTeX
@misc{drobac2026slidingwindow,
title = {{Sliding-Window Signatures as Features for Time Series Forecasting}},
author = {Nina Drobac and Margaux Brégère and Joseph de Vilmarest and Olivier Wintenberger},
howpublished = {HAL},
year = {2026},
month = oct,
eprint = {2510.12337},
archivePrefix = {arXiv},
url = {https://hal.science/hal-05308724v2},
note = {CC BY-NC 4.0},
}
RIS
TY - UNPB
TI - Sliding-Window Signatures as Features for Time Series Forecasting
AU - Nina Drobac
AU - Margaux Brégère
AU - Joseph de Vilmarest
AU - Olivier Wintenberger
PY - 2026
DA - 2026/10/26
PB - HAL
UR - https://hal.science/hal-05308724v2
AB - The evolution of a time series is often influenced not only by its past values but also by external covariates, giving rise to forecasting with exogenous variables. Nonlinear and delayed effects between covariates and the target series make this task particularly challenging. We propose a forecasting framework based on Ridge regression with sliding-window signature features, combining the computational efficiency of (semi-)parametric models with the flexibility of nonparametric approaches. These features capture complex dependencies between the target and the exogenous variables without relying on prior assumptions about the form of these relationships or the influence of past observations. We establish theoretical guarantees in the discrete-time setting, including universality of approximation and stationarity for the sliding-window transform. To ease implementation, we introduce a novel sequential algorithm that exploits window overlap for efficient signature updates. The method is evaluated on both synthetic and real electricity demand data. Results show that signature features effectively encode temporal and nonlinear dependencies, yielding accurate forecasts competitive with those based on expert knowledge and recent neural network approaches.
ER -