A Novel Evaluation Metric for Unsupervised Learning in AIS-Based Maritime Anomaly Detection: MADQI

In three linesNovel MADQI evaluation metric for unsupervised anomaly detection in maritime AIS datasets. Combines four indices (ARC, PPS, SDS, ECE) through automatic normalization. Achieves MADQI score of 80.37% on AIS data, with ECE=0.907 and ARC=1.000 for detecting abnormal vessel behavior.

Summary generated by Claude — human-verified