Forensic G-Indexes (G1, G2, G3): Graded Authorship-Filtered Citation Metrics for Researchers Using AI Agent and Google Scholar
The paper describes an AI-agent-assisted workflow integrated with a locally hosted Ollama language model and presents a comparative dataset of seventeen researcher profiles across multiple fields. The reported benchmark for comparison includes total citations, h-index, i10-index, absolute G-Index values, normalized G-Index percentages, and independent rankings. The analysis examines differences between conventional citation accumulation and citation counts associated with leading author positions, including a case study of Ali Khademhosseini. It further discusses how discrepancies among raw and normalized indicators may be used to identify profiles for publication-level review, while distinguishing bibliometric anomalies from evidence of research-integrity violations. The G-Indexes are proposed as complementary descriptive and screening indicators, not as direct measures of individual contribution or replacements for conventional metrics and qualitative research assessment.
Authors
- Daniel Saatchi
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-16
- DOI
- https://doi.org/10.5281/zenodo.24283616
- Primary Topic
- scientometrics and bibliometrics research
- Type
- preprint