Hasan Kurban

—

Unclaimed · AI-structured profileORCID OpenAlex

ML researcher advancing rigorous time-series forecasting evaluation and Arabic multimodal benchmarks

papers here

reviewed by Evolve

first-author

corresponding

—

works · OpenAlex

—

h-index · OpenAlex

Bibliometric figures come from OpenAlex (2026-10-17) 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—

Publication-output data is not yet available for this profile. This axis would measure the volume and cadence of scholarly output — total papers, venue types, and authorship roles — none of which are present in the supplied bibliometric snapshot, so no score is given.

Impact—

Impact cannot be scored yet: no citation counts, h-index, or downstream-usage indicators were provided. This axis would measure how widely the work is cited and taken up by other researchers, based on citations to the cataloged papers.

Focus100/100

Focus scores 100: the available field distribution assigns all classified work to a single field, machine learning (time-series forecasting), with no scattering across disciplines. The base is narrow — the field label currently covers one of two cataloged papers, the second being unclassified — so the perfect score should be read as concentrated evidence rather than a long-term field history.

Collaboration56/100

Collaboration scores 56, grounded in 16 distinct coauthors in the bibliometric record and a frequent-coauthor list of 12 names, including Samir Abdaljalil, Nadir Durrani, Fahim Dalvi, Erchin Serpedin, Mustafa Jarrar, and Shammur Absar Chowdhury. On the two cataloged papers Kurban is last author on one and middle author on the other, consistent with group work; the moderate score suggests the network is still growing.

Activity—

Activity data is insufficient for scoring. This axis would measure recent publishing cadence — papers per year and sustained engagement over time. Although both cataloged papers are dated 2026, the snapshot lacks the multi-year record this dimension requires, so no estimate is offered.

Reproducibility100/100

Reproducibility scores 100, the mean over exactly one assessed paper (reproPapers 1): 'When Does Retrieval Help Time-Series Forecasting?', which Evolve rated 'high' for reproducibility readiness. The paper's honest evaluation stance — reporting caveats on the diagnostic significance of its findings — aligns with that rating. With a single-paper base, the perfect score reflects strong practices on the work assessed but does not yet generalize across the catalog.

Research summary

Hasan Kurban (ORCID 0000-0003-3142-2866) is a machine-learning researcher whose cataloged work centers on evaluation methodology for time-series forecasting and on culturally grounded multimodal benchmarking. In Evolve's reviewed paper 'When Does Retrieval Help Time-Series Forecasting?' (arXiv, 2026), on which Kurban is the last-listed author, a rigorous and honest evaluation shows that the benefit of retrieval-augmented forecasting hinges on the window-to-period ratio rather than on the retrieval mechanism itself, with the analysis framed through seasonality, phase starvation, and spectral concentration. That paper carries Evolve's 'high' reproducibility-readiness rating, and the underlying metric (a mean of 100 across one assessed paper) yields a reproducibility dimension score of 100. A second 2026 arXiv paper, 'ImageEval 2026: Culturally Grounded Arabic Multimodal Evaluation,' extends the evaluation-first orientation to Arabic multimodal settings, with Kurban in a middle author position. The radar records 16 distinct coauthors and a collaboration score of 56; frequent collaborators include Nadir Durrani, Fahim Dalvi, Erchin Serpedin, Mustafa Jarrar, and Shammur Absar Chowdhury. Output, impact, and activity dimensions remain unscored because the underlying bibliometric counts are not yet available. This is an early-stage, name-resolved profile built from a limited bibliometric snapshot plus Evolve's reviews.

Collaboration pattern

Kurban works in team-based settings: the bibliometric record lists 16 distinct coauthors, and the frequent-collaborator list includes Samir Abdaljalil, Hunzalah Hassan Bhatti, Ahlam Bashiti, Farina Amir, Md Arid Hasan, Basel Mousi, Nadir Durrani, Fahim Dalvi, Zien Sheikh Ali, Erchin Serpedin, Mustafa Jarrar, and Shammur Absar Chowdhury. Authorship positions on the two cataloged papers (last on the time-series paper, middle on the Arabic multimodal paper) indicate contribution within larger multi-author groups spanning forecasting and multimodal evaluation.

Research keywords

time-series forecastingretrieval-augmented forecastingevaluation protocolseasonalityphase starvationspectral concentrationwindow-to-period ratioArabic multimodal evaluationcultural groundingreproducibility

Highlights

Retrieval helps forecasting only under the right window-to-period conditions
In the Evolve-reviewed 2026 arXiv paper (last-listed author), the benefit of retrieval-augmented time-series forecasting is shown to hinge on the window-to-period ratio rather than on the retrieval mechanism itself. The verdict calls the evaluation rigorous and honest, with caveats on the diagnostic significance of the findings.
Diagnostic framing: seasonality, phase starvation, spectral concentration
The same paper frames when retrieval pays off through concrete time-series diagnostics — seasonality, phase starvation, and spectral concentration — and carries a high reproducibility-readiness rating, anchoring the radar's reproducibility score of 100.
Co-author on culturally grounded Arabic multimodal evaluation
In 'ImageEval 2026: Culturally Grounded Arabic Multimodal Evaluation' (arXiv, 2026, middle author position), Kurban contributes to evaluation work for Arabic multimodal settings, extending the evaluation-first orientation beyond time series. This paper is not yet Evolve-reviewed and has no catalog slug.

Timeline

2026Last-listed author on 'When Does Retrieval Help Time-Series Forecasting?' (arXiv), showing that the retrieval benefit hinges on the window-to-period ratio.
2026Middle-position co-author on 'ImageEval 2026: Culturally Grounded Arabic Multimodal Evaluation' (arXiv).
2026The time-series paper enters the Evolve catalog with a 'high' reproducibility-readiness rating, anchoring the reproducibility score of 100.

Papers in the Evolve catalog

2026-09-18
When Does Retrieval Help Time-Series Forecasting?

Frequent co-authors

Samir Abdaljalil×1 Hunzalah Hassan Bhatti×1 Ahlam Bashiti×1 Farina Amir×1 Md Arid Hasan×1 Basel Mousi×1 Nadir Durrani×1 Fahim Dalvi×1 Zien Sheikh Ali×1 Erchin Serpedin×1 Mustafa Jarrar×1 Shammur Absar Chowdhury×1

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.

Dimension scores are computed from a limited public bibliometric snapshot; full publication and citation records are not reflected here.
Institution and country are not recorded in the source data; identity is anchored to ORCID 0000-0003-3142-2866, and the profile is name-resolved from public bibliographic data.
Output, impact, and activity axes are unscored because citation and publication-volume statistics are absent from the input data.
The reproducibility score (100) rests on a single assessed paper, and the second cataloged paper lacks field, verdict, and reproducibility metadata.