Reproducibility archive for Beyond Inventory Fit: Gosu-myeon dolmen prediction

This archive contains the source-transcription data, derived terrain predictors, analysis scripts and computational outputs supporting the study “Beyond Inventory Fit: Testing GIS-Based Dolmen Prediction with Newly Documented Monuments in Gosu-myeon, Gochang, Korea.” The archaeological dataset was reconstructed from the published Gosu-myeon catalogue by Noh and Lee (2024). It contains 97 coordinate-bearing management records at 91 distinct printed coordinate pairs, distributed across 25 archaeological groups. The source categories comprise 82 records not explicitly labelled new, nine additions within previously documented groups, and six records belonging to four new groups. Management records, coordinate locations, raster-cell observations and group centroids are distinguished in the accompanying crosswalks and documentation. The archive includes page-referenced source tables, record-to-location crosswalks, category assignments, six terrain predictors derived from CGIAR–CIAT SRTM v4.1, and Python scripts for three maximum-entropy model formulations. It also includes spatial cross-validation, matched AUC and area-based capture metrics, permutation importance, simple slope and proximity comparators, sensitivity analyses, fitted coefficients, prediction surfaces, software-environment records and execution logs. A contextual Gosu-myeon administrative-boundary snapshot, acquisition records, geometry checks and figure-generation code are included. The analysis is retrospective. The catalogue’s designation of a record as “new” does not independently establish its first-discovery date. Computational checks do not constitute independent field validation or legal certification of administrative boundaries. The full published catalogue and original SRTM tile are not redistributed. Third-party source notices and reuse conditions remain applicable. Funding: This research was supported by Konkuk University in 2026 and by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2023S1A5C2A02095114). AI-use disclosure: ChatGPT (OpenAI) assisted with English-language editing, documentation, code writing and revision, computational execution and result organisation, and table and figure preparation. Verification included source-page reconciliation, input-integrity and raster-cell checks, software tests, repeated model executions, independent metric calculations and sensitivity analyses. The author retains responsibility for the interpretation and accuracy of the deposited materials.