AnomaliX Detect

Anomaly detection and repair workbench (statistical + learning dual engines)

Product category: Anomaly detection / Cleaning

Applicable platforms: Python, Docker, Web UI

Technical affiliation: Z-score, IQR, IsolationForest, AutoEncoder

Programming language affiliation: Python

Tags: #AnomalyDetection #DataCleaning Product

Type: Anomaly detection / Cleaning

AnomaliX Detect provides statistical methods (quantile/robust scale) and representation learning (isolation forest, autoencoder) two types of detectors, covering patterns such as jumps, flat tops, gradual drifts, gaps, cycle misalignments; supports semi-automatic repair (forward filling/spline/similar day replacement) and human-machine collaborative review. For time series/multimodal training data purification, it can significantly reduce noise interference on parameter learning; for decision/replay data, it can label "unbelievable tracks" to avoid erroneous reward amplification. Engineering provides batch/stream two running modes, replayable audit tracks, and indicator panels. Anomalies are defined, detected, repaired, and replayed, forming a closed loop. 

Delivery method: Instant digital download after purchase
License: Single-user commercial license
Usage limit: One-time use
Support: Technical documentation provided in the delivery file; no human technical support included