Figure 4
The confusion matrix for the anomaly detection model consists of a 2x2 grid. The rows represent the actual classes, labeled as 0 and 1, while the columns represent the predicted classes, also labeled as 0 and 1. The matrix shows the following values: 4 true negatives for class 0, 1 false positive for class 0, 0 false negatives for class 1, and 11 true positives for class 1. The color gradient indicates the number of instances, with darker shades representing higher counts. The model demonstrates high accuracy in identifying anomalies, with a notable emphasis on recall to ensure no actual anomalies are missed.Confusion matrix of the XGBoost-based anomaly detection model. The model achieved 93.75% overall accuracy with 100% anomaly recall, correctly identifying all anomalous machining conditions while producing only one false positive. These results demonstrate the robustness of the proposed Physics-Guided Unified Damage Index framework for predictive maintenance and intelligent machining applications