Replication Package for "Process Mining and Stochastic Modeling for Software Process Forecasting: A Systematic Mapping Study with an LLM-Assisted Protocol"

Replication artefact for a systematic mapping study at the intersection of process mining, stochastic modeling, and machine-learning forecasting applied to software-development processes (Information and Software Technology, round-1 revision). Search: five databases (Scopus, IEEE Xplore, ACM Digital Library, SpringerLink, Web of Science); eligibility window January 1994–December 2025; searches executed 12 April 2026. After cross-tier deduplication the confirmed set is 340 distinct primary studies (404 raw combined includes; 318-study analytical subset). The package includes: search strings; deduplication, enrichment, and screening pipelines; LLM screening prompts and JSON schemas (claude-haiku-4-5 primary, claude-sonnet-4-6 verifier); 8-criterion reporting-quality rubric (Dybå and Dingsøyr, 2008); cross-model Cohen's kappa; blind human double-screening sheets and human–LLM kappa (T/A n=468, FT n=177); the 340-row included-study list (included_studies_340.csv); PRISMA snapshot; SEGRESS checklist; EC5 audit; sensitivity analysis. PDFs of primary studies are not redistributed (copyright); DOIs are in results/final_review/included_studies_340.csv.