AI-based Identity Fraud Detection: A Systematic Review

Dataset: AI-Based Identity Fraud Detection — Systematic Literature Review

This dataset accompanies the systematic literature review reported in AI-Based Identity Fraud Detection: A Systematic Review. It provides the screening logs for the title–abstract stage, the bibliographic sets used for full-text retrieval and assessment, and the full extraction dataset for all included studies. No full-text PDFs are included.

Contents

1. Rayyan Raw Screening Data (ZIP)

Export used for the title–abstract screening stage.

articles.csv — Deduplicated records retained for screening (1,119 unique articles).

customizations_log.csv — Rayyan action log capturing screening decisions (Include / Exclude / Maybe). CSVs are UTF-8, comma-delimited.

2. EndNote – Included Set (CSV)

CSV exported from EndNote containing the 193 records in the Rayyan Included set, used for full-text retrieval.

3. EndNote Library for Full-Text Stage (.enlx)

Archived EndNote library representing the reference set used at full-text screening.

Records are tagged and grouped into Included and Excluded papers.

No attached PDFs are included. Opening the .enlx restores an .enl file with its Data folder.

4. SLR Complete Dataset (SLR_Complete_Dataset.xlsx) (new)

Structured extraction dataset covering all 43 studies included in the final review. Contains five sheets:

README — Field definitions, screening summary, and citation guidance.

PRISMA_Summary — Aggregate record counts at each filter stage (Filters 1–5).

Included_Studies — Bibliographic information for all 43 included studies (title, authors, year, journal/conference, DOI, keywords, abstract).

Extraction_Data — Full coding of each included study, covering AI technology, task type, detection approach, applied context, fraud type, models, techniques, datasets used, evaluation metrics, findings, limitations, and future directions.

Quality_Scores — Quality assessment scores for all 43 studies across six criteria (Purpose, Relevance, Presentation, Evaluation, Findings, Directions); maximum score = 6; inclusion threshold ≥ 3.

5. PRISMA Flowchart and Checklist

PRISMA 2020 flow diagram documenting the full study selection process, from initial database searches through to final inclusion.

Access and Requests

Access to the live Rayyan workspace can be granted to editors and reviewers upon request during peer review. If additional formats are required, please contact the corresponding author.