Reservoir computing for macroeconomic forecasting with mixed-frequency data

Резервуарные вычисления для макроэкономического прогнозирования на основе данных смешанной частотности
Giovanni Ballarin, Πέτρος Δελλαπόρτας, Lyudmila Grigoryeva, Marcel Hirt, Sophie van Huellen, Juan‐Pablo Ortega
2023-12-07

U.S. GDP growthmacroeconomic forecastingmixed-frequency datamulti-frequency echo state networkreservoir computing
Macroeconomic forecasting has recently started embracing techniques that can deal with large-scale datasets and series with unequal release periods. Mixed-data sampling (MIDAS) and dynamic factor models (DFMs) are the two main state-of-the-art approaches to modeling series with non-homogeneous frequencies. We introduce a new framework, called the multi-frequency echo state network (MFESN), based on a relatively novel machine learning paradigm called reservoir computing. Echo state networks (ESNs) are recurrent neural networks formulated as nonlinear state-space systems with random state coefficients where only the observation map is subject to estimation. MFESNs are considerably more efficient than DFMs and can incorporate many series, as opposed to MIDAS models, which are prone to the curse of dimensionality. All methods are compared in extensive multistep forecasting exercises targeting U.S. GDP growth. We find that our MFESN models achieve superior or comparable performance over MIDAS and DFMs at a much lower computational cost.
1
Compared with dynamic factor models, MFESNs are considerably more computationally efficient and can handle large numbers of time series.
2
In extensive multistep U.S. GDP growth forecasting exercises, MFESNs achieve superior or comparable accuracy to MIDAS and dynamic factor models at substantially lower computational cost.
3
MFESNs avoid the dimensionality limitations affecting MIDAS models while incorporating many mixed-frequency series.
4
MFESNs represent echo state networks as nonlinear state-space systems with random recurrent coefficients, estimating only the observation map.
5
The paper introduces multi-frequency echo state networks (MFESNs), a reservoir-computing framework for forecasting macroeconomic data with unequal release frequencies.

U.S. GDP growth with mixed-frequency macroeconomic time series

Multistep forecasting performance and computational efficiency of reservoir computing compared with MIDAS and dynamic factor models

Publication Details
Publication Date
2023-12-07
Journal
Publisher
ISSN
Cited by
31
Access Type
Author Information
Authors
Giovanni Ballarin
Πέτρος Δελλαπόρτας
Lyudmila Grigoryeva
Marcel Hirt
Sophie van Huellen
Juan‐Pablo Ortega
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%