Early-career ML researcher at Rajshahi spanning computational gait analysis, topic modeling, and applied deep learning
2
papers here
1
reviewed by Evolve
0
first-author
0
corresponding
—
works · OpenAlex
1
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—
Not scored: the snapshot provides no aggregate output metrics beyond recent-year counts, so total publications, venue types, and full publication history are unavailable. This axis would measure cumulative and per-year production volume and its composition across venues.
Impact—
Not scored: citation-based impact data is insufficient — the only signals available are an h-index of 1 and an i10-index of 0, far too early to indicate influence. This axis would measure citation counts, citing works, and field-normalized citation impact.
Focus100/100
Scored 100 on Evolve's focus axis: the OpenAlex field distribution contains exactly one labeled field (biomedical machine learning / computational gait analysis), and Topic Modeling is the most frequent of five recorded topics (2 versus 1 each for neural network applications, gait/falls prevention, media studies, and software engineering). Within the limited record, work clusters tightly rather than scattering.
Collaboration44/100
Scored 44 on collaboration: the recent three-year window records 8 distinct coauthors, with recurring partners including Jungpil Shin, Abu Saleh Musa Miah, Md. Sifat, and Sania Akter. Both Evolve-catalog papers list Islam as a middle author rather than corresponding author, consistent with a contributor role inside small team projects and moderate rather than dense network reach.
Activity67/100
Scored 67 on activity: OpenAlex counts 9 works in the recent three-year window, with countsByYear showing a sharp ramp from 1 work in 2025 to 8 in 2026. The pace is high for an early-stage record but concentrates in a single burst year, so sustained output across multiple years is not yet demonstrated.
Reproducibility55/100
Scored 55 on reproducibility, computed from 1 reviewed paper (reproPapers = 1, mean 55). That paper — the Parkinson's disease COP-representation study — carries medium reproducibility readiness: Evolve's verdict notes nested cross-validation performed within a single database, with external validation absent, so claims are supported internally but not yet tested outside it.
Research summary
Akif Islam is an early-career researcher at the University of Rajshahi, Bangladesh, whose OpenAlex-resolved record is young but fast-moving. The snapshot counts 9 works in the recent three-year window, jumping from 1 work in 2025 to 8 in 2026, with 8 distinct coauthors. Topic Modeling is the most frequent OpenAlex topic (2 papers), alongside single appearances in advanced neural network applications, balance/gait/falls prevention, media studies, and software engineering; the labeled field distribution contains one field, biomedical machine learning / computational gait analysis, which underlies Evolve's focus score of 100. The Evolve catalog holds two 2026 arXiv papers with Islam as a middle author: a compact stance-indexed anterior-posterior center-of-pressure representation for Parkinson's disease classification from plantar vertical ground reaction force (reviewed — nested cross-validation supports the claim within one database, external validation absent, medium reproducibility readiness, reproducibility mean 55), and PermitGPT, a generative-AI pipeline for construction hazard forecasting, permit prediction, and community impact (not yet reviewed). Citation indicators remain minimal (h-index 1, i10-index 0), so the output and impact axes are unscored. Taken together, this is an early-stage, collaboration-driven profile concentrated in 2026.
Collaboration pattern
Islam's recent record shows 8 distinct coauthors in the OpenAlex three-year window (collaboration score 44). Recurring partners include Jungpil Shin, Abu Saleh Musa Miah, Md. Sifat, Sania Akter, Md. Ekramul Hamid, Farjana Aktar, Mohd Ruhul Ameen, and Momen Khandoker Ope. In both catalog papers Islam appears as a middle author and not corresponding author, indicating a team-contributor pattern in multi-author arXiv projects spanning gait analysis and generative AI rather than solo or led outputs.
Research keywords
Highlights
Compact stance-indexed AP-COP representation for Parkinson's gait classification
PermitGPT: generative-AI pipeline for construction hazard and permit prediction
Second 2026 arXiv catalog entry, middle-author. Proposes a unified generative-AI pipeline covering construction hazard forecasting, permit prediction, and community impact assessment; Evolve has not yet published a verdict, keywords, or reproducibility rating for it.
Timeline
- 2025First recorded OpenAlex output: 1 work, marking the start of the indexed publication record.
- 2026Co-authored (middle position) the arXiv paper on a compact stance-indexed anterior-posterior COP representation for Parkinson's disease classification from plantar VGRF.
- 2026Co-authored (middle position) the PermitGPT arXiv paper, a generative-AI pipeline for construction hazard forecasting, permit prediction, and community impact.
- 2026Output peak: 8 works recorded in OpenAlex for the year, spread across five topics led by topic modeling.
Papers in the Evolve catalog
Frequent co-authors
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.
- Bibliometric snapshot is OpenAlex-only; no ORCID is linked, so the record for the name 'Akif Islam' is name-resolved from public bibliographic data and may include identically named authors.
- Citation impact is too early-stage to score (h-index 1, i10-index 0); the output and impact axes are left unscored rather than estimated.
- Only one catalog paper carries a full Evolve review (verdict, keywords, reproducibility); PermitGPT is listed without review data, and the reproducibility score rests on a single assessed paper.
- Field distribution and topic counts reflect the small recent record (9 works in 3 years) and will shift as more works are indexed.