Anomaly Detection

Control Chart Analysis and Anomaly Detection

Combine Xbar-R / Xbar-S, EWMA, CUSUM, Z-score, IQR, wafer map distribution bias, and correlation outlier detection to support early detection of process and equipment abnormalities.

Xbar-R control chart: switch between Shewhart, EWMA, and CUSUM views to monitor time-series drift
Wafer control chart: monitor statistics for each split region over time
Correlation outlier detection: extract data outside correlation trends as anomaly candidates

Statistical Anomaly Detection

Detect outliers from measurement distributions using Z-score and IQR. Also effective for removing abnormal data during new product ramp-up.

Control Chart Analysis

Monitor process variation, drift, and signs of equipment abnormalities over time using Xbar-R / Xbar-S, EWMA, and CUSUM.

Spatial and Correlation Analysis

Extract chips outside wafer map distributions and correlation trends to help identify local anomalies and influential parameters.

Automation

Automation Features

Improve daily monitoring and reporting efficiency through automatic analysis for new lots, anomaly alerts, and periodic report generation.

Automatic analysis for new lot input
Alerts when anomalies are detected
Periodic report generation

Next

Review managed data structures and performance

Review technical specifications including lot-level variance, STDF/CSV/JSON support, and MySQL 256-sharding.

View Technical Specifications