WIRED: Weighted Adaptive Prediction with Structured Dependence for Probabilistic Multiseries Forecasting

📅 2026-08-13
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
This work addresses the performance bottleneck in joint probabilistic forecasting of multivariate time series arising from the entangled modeling of marginal distributions and cross-series dependencies. To resolve this, the authors propose WIRED, a novel approach that decouples adaptive marginal prediction via expert aggregation from dependency structure reconstruction based on copulas. Specifically, marginal forecasts are generated using a CRPS-weighted adaptive ensemble, while dependencies are captured through Gaussian or Student-t copulas. The framework is rigorously evaluated via rolling-origin validation and ablation studies on both synthetic data and the EuStockMarkets benchmark. Results demonstrate the efficacy of the proposed architecture and reveal limitations in current CRPS-weighting strategies, showing that simple equal-weighting or bootstrap aggregation remains competitive for marginal modeling—offering new insights for designing regularized ensembles in probabilistic forecasting.
📝 Abstract
This paper presents WIRED, an R package algorithm for joint probabilistic forecasting of multiple related time series. WIRED combines a library of simple marginal predictive distributions, CRPS-based adaptive mixture weights, and a Gaussian or Student t copula for cross-series simulation. We evaluate the implementation in a benchmark with four synthetic data-generating processes (DGPs), three forecast horizons, 30 replicates per DGP-horizon pair, nine ablations and external baselines, and a rolling-origin study on the built-in EuStockMarkets data. The central contribution is architectural and diagnostic. WIRED separates adaptive marginal expert aggregation from dependence reconstruction; the benchmark supports explicit dependence modeling, but shows that the current CRPS-extrapolated softmax weighting is not yet robust enough to dominate simpler bootstrap or equal-weight alternatives. The paper therefore identifies a working layer of the design, a bottleneck in the marginal aggregation layer, and a concrete research path for more regularized probabilistic ensemble construction.
Problem

Research questions and friction points this paper is trying to address.

probabilistic forecasting
multiseries
dependence modeling
adaptive weighting
marginal aggregation
Innovation

Methods, ideas, or system contributions that make the work stand out.

probabilistic multiseries forecasting
adaptive mixture weighting
copula-based dependence modeling
CRPS-based ensemble
modular forecast architecture