Synthetic Reproducibility Package for "Reducing Credential Exposure Risk in UK SMEs: A Design Science Evaluation of Contextual Machine Learning and Compliance Automation"

This repository provides a synthetic reproducibility package supporting the machine learning pipeline described in the manuscript: “Reducing Credential Exposure Risk in UK SMEs: A Design Science Evaluation of Contextual Machine Learning and Compliance Automation.” The package includes: Synthetic feature representations Random Forest model weights and configuration (.joblib) Feature importance scores Training and evaluation scripts The materials enable replication of the computational workflow and evaluation methodology described in the manuscript. No operational SME-derived data, raw credential artefacts, organisational metadata, or field engagement data are included; all materials are synthetic and provided solely for methodological transparency.