arXiv subjects

Giancarlo Vercellino

Publications and source records attributed to Giancarlo Vercellino.

4 recordsLinked to original sources

arXiv2026

TEMPER: Temporal Encoder-Masked Probabilistic Ensemble Regressor for Time-Series Forecasting

Probabilistic forecasting requires accurate central predictions and calibrated uncertainty estimates. This paper presents TEMPER, the Temporal Encoder-Masked Probabilistic Ensemble Regressor, a univariate time-series forecasting algorithm that combines a temporal autoencoder, a differentiable masked neural decision forest, continuous ranked probability score (CRPS) training, and Gaussian-mixture post-processing. The R implementation is built on torch for R and returns horizon-wise density, distribution, quantile, and sampler functions. We evaluate TEMPER on three deterministic synthetic level series with trend, periodic, regime-switching, nonlinear-threshold, and heteroskedastic components. Across 96 rolling-origin forecasts at horizons t + 1, t + 5, t + 20, and t + 60, TEMPER obtains 2.824% mean CRPS normalized by origin level, 3.635% median absolute error, and 68.8% empirical 90% interval coverage after training with a 300-epoch cap and early-stopping patience of 100. A naive persistence bootstrap has the best aggregate CRPS, 2.763%, while TEMPER has the best median absolute error and the best CRPS at t+1 and t+5. The ablation study uses matched series-origin-horizon cells, horizon-wise CRPS deltas, endpoint sensitivity summaries, and a calibration-specific interval study. Relaxing the learned mask improves average CRPS by 0.472 percentage points on the ablation subset, mainly through long-horizon gains. A twofold interval inflation improves held-out coverage from 54.2% to 91.7% and gives the best 90% interval score among tested calibration rules. The results identify calibration, horizon-specific tuning, and component selection as the central research priorities.

stat.ML↗

arXiv2026

Rollcast: Proper-Score Gated Rolling Anchors for Adaptive Probabilistic Time-Series Forecasting

Rollcast is a probabilistic forecasting method for univariate time series that combines a compact set of rolling statistical anchors rather than relying on a single global model. Rolling means, medians, extrema, regression endpoints, and quantiles define candidate forecast locations and a representation of the current state. A state-dependent softmax gate learns anchor probabilities by minimizing negative log predictive density, while residual distributions retrieved from similar historical states provide local uncertainty. Recursive simulation propagates the resulting mixture over multiple forecast horizons. The method is evaluated in a Monte Carlo study covering eight data-generating processes, including autoregressive, random-walk, local-trend, threshold, regime-switching, stochastic-volatility, heavy-tailed, and variance-break dynamics. Across 2,000 independent fitted series, Rollcast is compared with the true conditional predictive distribution generated by an oracle simulator. Overall empirical coverage is 86.2% for nominal 90% intervals and 91.5% for nominal 95% intervals. Predictive intervals are on average 13.6% wider than the oracle at the 90% level, while CRPS is 14.4% higher than oracle CRPS. Performance is closest to the oracle under autoregressive, threshold, stochastic-volatility, heavy-tailed, and variance-break dynamics, while local trends and regime switching are more challenging. The results indicate that Rollcast can construct competitive probabilistic forecasts from simple, interpretable local summaries, while also identifying limitations in calibration and recursive uncertainty propagation.

stat.ML↗

arXiv2026

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

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.

stat.ME↗

arXiv2026

Conditional Regime Analog Forecasting with Trajectories: A Nonparametric Framework for Multivariate Probabilistic

We propose Conditional Regime Analog Forecasting with Trajectories (CRAFT), a nonparametric framework for multivariate probabilistic time-series prediction. The method constructs paired backward and forward trajectory profiles from cumulative multi-horizon returns, learns recurrent low-dimensional regime labels in both spaces using singular value decomposition, change-point segmentation, and segment clustering, and estimates a backward-to-forward conditional regime correspondence. Forecast distributions are obtained by sampling historically realized future trajectory profiles according to a composite compatibility score that combines future-regime correspondence with similarity to the current backward trajectory profile. Unlike parametric vector autoregressions or Gaussian state-space models, CRAFT preserves empirical cross-sectional and multi-horizon dependence by resampling complete future paths. We describe the estimator, its diagnostics, and a reproducible simulation benchmark comparing CRAFT with direct analog resampling, SVD analogs, unconditional bootstrap, OLS VAR bootstrap, and random-forest forecasts.

stat.ME↗