SHAD
Scality High-dimensional Anomaly Detection benchmark
A fully annotated end-to-end benchmark for time series anomaly detection containing 215 high-dimensional multivariate real-world cloud storage datasets with rich semantic labels.
How to run
$ npm i -g lettokenburn-cli$ ltb bench info shad
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
2 sourced entriesEvery number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsAbout
Content & Construction: SHAD is a dataset built from the Scality RING distributed storage platform, comprising 215 high-dimensional time series. It features curated anomalies affecting different architectural layers and includes per-dimension annotations mapping sensors to system components.
Scoring & Key Findings: The benchmark evaluates detection accuracy across various detectors, spatial explainability via dimension contribution analysis, and interpretability using frozen LLM baselines for anomaly localization.