Filtering and Smoothing with Score-Driven Models
Abstract
Score-driven models are, by construction, purely predictive filters. When such models are read as filters for an underlying latent process, rather than as data generating processes, this is a restriction rather than a feature, because the contemporaneous and the future observations also carry information on the current state. Starting from the observation that the Kalman filter and smoother recursions for linear Gaussian models can be written in terms of the score of the conditional log-likelihood, and that the predictive step then takes the form of a score-driven recursion, we generalize score-driven models along this direction by deriving an update filter and a smoother, which exploit the contemporaneous observation and the whole sample respectively, together with the corresponding conditional variances. In extensive Monte Carlo analyses the update filter lowers the mean square error of the predictive filter by 3.5% to 8.8%, and the smoother by 32% to 44%, with the ordering holding in every replication; confidence bands built on the associated conditional variances attain their nominal coverage, whereas bands that ignore filtering uncertainty capture less than a third of it. Empirically, we demonstrate the benefits of employing score-driven models as filters rather than as purely predictive processes, showing that the resulting smoothed estimates align more closely with realized quantities than their predictive counterparts.
Related
- Kalman Filtering and Smoothing for Improving Precision in Horvitz--Thompson Estimation of Infectious Disease Prevalence
- Covariance-Informed Ensemble Score Filtering: Embedding Localized Covariance Geometry into Score-Based Transport
- AECSF: Adaptive Ensemble Conditional Score Filtering for High-Dimensional Nonlinear Data Assimilation
- A Survey of Nonlinear Filtering Methods for Data Assimilation and Recent Advances in Generative AI-Enabled Score Filters
- Tweedie's Formula and Score-Driven Updating
BibTeX
@misc{buccheri2026filtering,
title = {{Filtering and Smoothing with Score-Driven Models}},
author = {Giuseppe Buccheri and Giacomo Bormetti and Fulvio Corsi and Fabrizio Lillo},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.24571},
eprint = {2609.24571},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.24571},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Filtering and Smoothing with Score-Driven Models
AU - Giuseppe Buccheri
AU - Giacomo Bormetti
AU - Fulvio Corsi
AU - Fabrizio Lillo
PY - 2026
DA - 2026/09/21
PB - arXiv
DO - 10.48550/arxiv.2609.24571
UR - https://arxiv.org/abs/2609.24571
AB - Score-driven models are, by construction, purely predictive filters. When such models are read as filters for an underlying latent process, rather than as data generating processes, this is a restriction rather than a feature, because the contemporaneous and the future observations also carry information on the current state. Starting from the observation that the Kalman filter and smoother recursions for linear Gaussian models can be written in terms of the score of the conditional log-likelihood, and that the predictive step then takes the form of a score-driven recursion, we generalize score-driven models along this direction by deriving an update filter and a smoother, which exploit the contemporaneous observation and the whole sample respectively, together with the corresponding conditional variances. In extensive Monte Carlo analyses the update filter lowers the mean square error of the predictive filter by 3.5% to 8.8%, and the smoother by 32% to 44%, with the ordering holding in every replication; confidence bands built on the associated conditional variances attain their nominal coverage, whereas bands that ignore filtering uncertainty capture less than a third of it. Empirically, we demonstrate the benefits of employing score-driven models as filters rather than as purely predictive processes, showing that the resulting smoothed estimates align more closely with realized quantities than their predictive counterparts.
ER -