Module diagnostics_ml 

ML-based Anomaly Detection and Forecasting

This module provides machine learning components for stack diagnostics:

• Isolation Forest for anomaly detection
• Error Forecaster for time series prediction

Toyota Way Principles

• Jidoka: ML anomaly detection surfaces issues automatically
• Genchi Genbutsu: Evidence-based diagnosis from actual data

Structs

ErrorForecaster

Simple exponential smoothing for time series forecasting

ForecastMetrics

Forecast error metrics

IsolationForest

Isolation Forest for anomaly detection Implements a simplified version of the algorithm from Liu et al. (2008)