Why
The PoC runs isolation-forest and statistical detectors on transaction data and reports precision/recall at several thresholds, because the ROC curve is the honest deliverable.
How it works
Not yet built.
Workspace Index › Dev Notes › Anomaly detection — the alert that fires on everything or nothing
#181PoC
Anomaly detectors flag points that do not fit a learned normal, and the entire product is the threshold: too tight and it is noise, too loose and it misses the one event that mattered.
The PoC runs isolation-forest and statistical detectors on transaction data and reports precision/recall at several thresholds, because the ROC curve is the honest deliverable.
Not yet built.
이상 탐지기는 학습된 정상에 맞지 않는 점을 표시하며, 제품의 전부는 임계값입니다: 너무 좁으면 잡음, 너무 넓으면 정작 중요한 그 한 사건을 놓칩니다.
이 PoC는 거래 데이터에 격리 숲과 통계적 탐지기를 돌려 여러 임계값에서 정밀도/재현율을 보고합니다. ROC 곡선이 정직한 산출물이기 때문입니다.
아직 만들지 않음.