STGAD: Anomaly Detection Performance Metrics for Multivariate Time Series

5.5 KB of Excel data compares the STGAD anomaly detection framework's performance against a baseline method. The dataset likely contains saliency overlap metrics for multivariate time series across five benchmark domains. It was authored by Xiao Liao and last updated on figshare in May 2026.

Use Cases

• Benchmarking anomaly detection methods based on saliency overlap metrics.
• Evaluating adversarial stability and temporal dependency modeling in generative frameworks.
• Calibrating anomaly scores across different datasets like SMD, SMAP, and SWaT.
• Analyzing the effectiveness of dual-score fusion and distribution-adaptive thresholding.

Strengths

• Data is licensed under CC-BY-4.0, allowing for open reuse.
• The description references five benchmark datasets (SMD, SMAP, MSL, SWaT, MIT-BIH), suggesting a multi-domain evaluation.
• The framework described employs a dual-score generative-adversarial approach with Transformer encoders.

Limitations

• The dataset is only 5.5 KB, indicating a very limited scope, likely containing summary metrics rather than raw time series.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Likely contains experimental results from the STGAD framework evaluation.

Freshness

Last updated 2026-05-21 17:42:24; freshness should be verified.