Anjum | 2026-nnn

**Repository**: Not yet mirrored into the codecheckers organisation. The code and data are archived at https://doi.org/10.5281/zenodo.22078287 (v3.1.3, CC-BY 4.0) and the analysis code alone is small. @nuest, could a repository be created for the mirror, or should I prepare it another way?

**Workflow**: "Evaluating ML-Based Anomaly Detection on Unified OpenTelemetry Telemetry: An Empirical Study Across Traces, Metrics, and Logs", IEEE Access, https://doi.org/10.1109/ACCESS.2026.3705430. The archive contains raw anomaly scores, labels, and analysis scripts. Running the prevalence-sensitivity script against the archived data regenerates the published prevalence tables. The README documents the commands and a pinned environment (numpy 2.5.2, pandas 3.0.5, scikit-learn 1.9.0, Python 3.14). A clean-environment re-execution on 2026-08-24 reproduced the archived tables exactly, so the check should be mechanical.

Start with the README and analysis scripts in the Zenodo archive, then run the prevalence-sensitivity script against the archived data in the pinned Python environment. Compare the regenerated prevalence tables with the published tables; done means the check is reproducible and the repository mirror path is confirmed.

Written by the indexing model from the issue text.