Interpretable Anomaly Detection for Numerical Data in Python Using PyNomaly
summary
PyNomaly implements LoOP to produce 0–1 probability scores for numerical anomalies and ships as a lean Python + numpy library with optional Numba and pandas support. It includes a streaming LoOP from 2016 and accepts custom distance matrices; the repo reports ~54,000 downloads in a recent month. Demos show it flags low-density points reliably but fails when anomalies sit inside normal local density or when anomalies are aggregate/windowed rather than per-sample.
AI-written summary · about
chapters (7)
- 0:00Interpretable anomaly detection
- 2:12PyNomaly and LOOP algorithm
- 7:07Why LOOP for anomaly detection
- 13:29LOOP limitations
- 15:37PyNomaly demo and usage
- 18:47Future outlook and contributions
- 21:18Q&A
related talks
28mPopulation Anomaly Detection with PyTorch
01h 16mAnomaly Detection
30mBetter data quality with ML & visualization
45mAnomaly Detection in Time Series
01h 53mAnalyze your data in Python
45mBuilding an Anomaly Detector for High-Load from Python and Duct Tape
55mDerivatives Analytics with Python & Numpy
28mHow Machine Learning Can Achieve Data Quality Monitoring At Scale