Data Sharing Policy

Data Sharing Policy

1. Principle

Results reported in IJCIE should be verifiable. The journal therefore encourages authors to make the data, code and other materials underlying their findings openly available, and requires every article to carry a data availability statement.

2. Data availability statement (required)

Every article must include a statement before the references, in one of these forms:

  • Openly available: "The data supporting the findings of this study are openly available in [repository] at [DOI or URL]."
  • Public dataset used: "This study analysed publicly available datasets, available at [URL/DOI]."
  • Available on request: "The data supporting the findings of this study are available from the corresponding author on reasonable request."
  • Restricted: "The data are not publicly available because [reason, for example participant privacy or a third-party agreement]. [Description of any conditional access.]"
  • No new data: "No new data were created or analysed in this study."

A statement that simply says "data available on request" is accepted, but authors who use it are expected to honour such requests; failure to do so may be reported under the misconduct policy.

3. Recommended practice

  • Deposit data in a recognised repository that issues a persistent identifier, such as Zenodo, figshare, Dryad, IEEE DataPort or a domain-specific archive.
  • Deposit code in a public version-controlled repository and archive a citable release.
  • Cite data and code in the reference list, with the DOI, exactly as any other source.
  • Include a README describing the file structure, formats, units, and the exact procedure to reproduce each reported table and figure.
  • License data and code openly, for example CC BY or CC0 for data and a recognised open source licence for code.

4. Machine learning specific requirements

For work that reports trained models, the manuscript must state the dataset and its version, the train/validation/test split protocol, whether the split is subject-independent where that matters, all hyperparameters, the random seed policy, the hardware used and the number of runs over which reported results are averaged, together with a dispersion measure. Results reported from a single run without this information may be returned before review.

5. Editorial verification

Editors and reviewers may request the underlying data or code during review as a condition of acceptance. Refusal without an adequate reason is grounds for rejection.

6. Personal, sensitive and third-party data

Data derived from human participants must be anonymised before sharing, and sharing must be consistent with the consent obtained and with applicable data protection law. Where sharing is not possible, the restriction and its reason must be stated. Third-party data may only be redistributed where the licence permits it; otherwise the source and access route must be described.

7. Retention

Authors are expected to retain the data underlying a published article for at least five years and to make it available to an editorial investigation on request.