Manufacturing Quality Anomaly Detection Dashboard

Proactive yield monitoring, anomaly prioritization, and generated pre/post action review for high-volume electronics manufacturing.

Quality Monitoring Overview

Yield Trend with Baseline Reference

Defect Pareto Analysis

Amber highlight — D042: largest defect-share increase

Equipment Concentration Analysis

Red highlight — EQP_07: suspicious equipment with highest failure-share delta

ML Anomaly Ranking

ML-assisted ranking method

Model: Isolation Forest
Input signals: Yield rate, fail rate, retest pass rate, rolling yield delta, defect-code concentration, equipment concentration
Purpose: Prioritize abnormal manufacturing patterns for engineering review through ML-assisted anomaly ranking and signal prioritization.

Limitation

The model does not automatically determine root cause. Corrective Action Effectiveness Review compares generated pre-action and post-action windows against deterministic improvement thresholds. The comparison summarizes post-action improvement signals and does not establish causal effect; action dates remain visible as context regardless of rank.

Rank	Date	Action Context	Anomaly Score	Yield Rate	Suspected Signal	Engineering Interpretation