mononitogoswami/tsad-model-selection

Code for "Unsupervised Model Selection for Time-series Anomaly Detection", ICLR 2023.

42
/ 100
Emerging

This project helps operations engineers and data scientists identify the best anomaly detection model for their time-series data without needing labeled examples. It takes in raw time-series datasets and outputs a ranking of anomaly detection models, indicating which one is most likely to perform best. This is especially useful for anyone dealing with vast amounts of sensor data, system logs, or financial streams where manual labeling is impractical.

No commits in the last 6 months.

Use this if you have time-series data and need to find the most effective anomaly detection model among many options, but lack the labeled anomaly data traditionally required for model selection.

Not ideal if you already have a method for generating labeled anomaly data or if you need to select models for non-time-series data.

operations-monitoring predictive-maintenance fraud-detection system-health time-series-analysis
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 17 / 25

How are scores calculated?

Stars

87

Forks

15

Language

Jupyter Notebook

License

Apache-2.0

Last pushed

Dec 14, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/mononitogoswami/tsad-model-selection"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.