Giancarlo Vercellino

—

Unclaimed · AI-structured profileOpenAlex

Probabilistic time-series forecasting; solo arXiv author of CRPS-based ensemble and rolling-anchor forecasters

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-16) and cover the author's entire public record, not just the papers listed here. Reproducibility signal comes from Evolve's own reviews.

Research profile

Output—

This axis measures publication volume and productivity over time, but the indexed record is too thin to score: the snapshot reports an h-index of 0 and an i10-index of 0, so no output score is estimated.

Impact—

This axis measures citation uptake and scholarly influence of published work. No usable impact data is available yet — h-index and i10-index are both recorded as 0 — so the axis is left unscored rather than estimated.

Focus0/100

The focus score is 0, while the raw field distribution is tightly bounded: fieldDist splits 0.5/0.5 between 'probabilistic time-series forecasting' and 'statistical machine learning (time-series forecasting)' — two adjacent labels covering the same domain. Topic counts reinforce this: time-series analysis and forecasting (15), forecasting techniques (10), stock-market methods (5), energy load (3), and R-based analysis (3). The profile is deep but narrow in scope.

Collaboration0/100

The collaboration score is 0, and the raw record explains why: 0 distinct co-authors are listed and the frequent-coauthor list is empty. Both reviewed 2026 arXiv papers name him as first author and corresponding author, consistent with a strictly solo working pattern across probabilistic forecasting projects. Any team-based or supervised work would not appear in this snapshot, so the picture reflects indexed metadata rather than a complete career record.

Activity77/100

The activity score of 77 rests on recent-year output: the snapshot counts 4 works in 2022, 1 in 2023, 1 in 2024, 5 in 2025, and 7 in 2026 — 13 works within the last three years. The trajectory dips through 2023–2024, then accelerates sharply in 2025–2026, coinciding with the two reviewed arXiv preprints in probabilistic forecasting. Posting cadence, not citations, drives this axis.

Reproducibility100/100

Reproducibility scores 100, the maximum, computed across the 2 reviewed papers with a reproducibility mean of 100; both TEMPER and Rollcast carry a 'high' reproducibility-readiness rating. The verdicts also note methodological honesty: TEMPER's synthetic evaluation shows a persistence baseline beating its CRPS, and Rollcast reports 86.2% empirical coverage against a 90% nominal level — transparent, if still early-stage, empirical practice.

Research summary

Giancarlo Vercellino is an early-stage profile built from OpenAlex bibliographic data: a researcher working almost entirely in probabilistic time-series forecasting and statistical machine learning, with the field split evenly (0.5/0.5) between those two adjacent labels. Topic counts concentrate on time-series analysis and forecasting (15 topics), forecasting techniques and applications (10), with smaller applied strands in stock-market forecasting (5), energy load and power forecasting (3), and data analysis with R (3). Output is recent and rising: the snapshot records 4 works in 2022, 1 in 2023, 1 in 2024, 5 in 2025, and 7 in 2026, with 13 works in the last three years. Citation-based indicators are not yet meaningful (h-index and i10-index both 0), and no co-authors are recorded. Evolve has reviewed two 2026 arXiv preprints, both with him as first and corresponding author: TEMPER, an autoencoder–neural-forest–CRPS ensemble regressor whose synthetic-only evaluation shows a persistence baseline beating its CRPS, and Rollcast, a proper-score-gated rolling-anchor forecaster reaching 86.2% empirical coverage against a 90% nominal target. Both rate high on reproducibility readiness (reproducibility mean 100 across 2 papers).

Collaboration pattern

The indexed record shows no collaboration: 0 distinct co-authors, an empty frequent-coauthor list, and both reviewed 2026 arXiv preprints carrying him as first and corresponding author. All current work in probabilistic forecasting appears self-contained; no lab, supervisor, or consortium is visible in this snapshot. For outreach, the only public touchpoints are the arXiv preprints themselves.

Research keywords

probabilistic forecastingCRPS (continuous ranked probability score)forecast combinationmixture of expertsproper scoring rulespredictive calibrationuncertainty quantificationneural decision forestsautoencodertime-series analysis and forecastingenergy load forecastingstock market forecasting

Highlights

TEMPER: an honest-but-immature probabilistic ensemble forecaster
Evolve's review describes TEMPER as an autoencoder + neural-forest + CRPS forecaster; on synthetic series a persistence baseline beats its CRPS, while every ablated component helps — honest, but not yet mature. It appeared on arXiv in 2026 with him as first and corresponding author, and carries a high reproducibility-readiness rating.
Rollcast: proper-score gated rolling anchors under synthetic-only evaluation
Rollcast gates rolling-stat anchors by log score. The review reports 86.2% empirical coverage against a 90% nominal target and close-to-oracle efficiency, but flags synthetic-only evaluation and the absence of practical baselines. arXiv 2026, first and corresponding author, high reproducibility readiness.
Perfect reproducibility signal across reviewed work
Both reviewed papers carry a 'high' reproducibility-readiness rating, yielding a reproducibility mean of 100 across 2 papers — the basis of the 100/100 reproducibility dimension. The signal reflects transparent, verifiable empirical setups rather than citation strength.

Timeline

2022First indexed burst: the OpenAlex snapshot records 4 works in 2022, the highest count before 2026.
2023Output slows to a single indexed work in 2023.
2024Again only 1 indexed work — a quiet year before the current acceleration.
2025Activity resumes: 5 indexed works, concentrated in forecasting-related topics.
2026Peak year so far: 7 indexed works, including the two reviewed arXiv preprints TEMPER and Rollcast, both first- and corresponding-authored.

Papers in the Evolve catalog

2026-09-20first authorcorresponding
TEMPER: Temporal Encoder-Masked Probabilistic Ensemble Regressor for Time-Series Forecasting
2026-09-03first authorcorresponding
Rollcast: Proper-Score Gated Rolling Anchors for Adaptive Probabilistic Time-Series Forecasting

About this profile

This profile is AI-structured by Evolve.Pub from public bibliographic data (OpenAlex) and Evolve's own paper reviews. It is not written or endorsed by the researcher unless the profile is claimed. Radar dimensions are computed from real bibliometrics with fixed formulas — raw figures are shown alongside; missing dimensions mean insufficient data, never zero. Errors? Claim the profile to correct or hide it.

Identity is name-resolved from public OpenAlex bibliographic data; no ORCID, institutional affiliation, or country is attached to this record.
All bibliometric indicators come from OpenAlex only; no other bibliographic sources are included.
Citation-based indicators are 0 (h-index and i10-index), so the output and impact dimensions are missing rather than merely low.
Only 2 indexed works have Evolve reviews, and both are 2026 arXiv preprints; the other works in the counts-by-year record are unreviewed.