INCLAIR: Longitudinal Clinical Anomaly Detection with Informed Reasoning
July 31, 2026
INCLAIR is a framework for detecting anomalies in sparse, unequal-length clinical profiles. It uses a complete mean subsequence score to aggregate evidence and generate grounded natural-language explanations while controlling combinatorial inference costs.
HOW THIS AFFECTS YOU
●
researcherThe use of U-statistics to control inference cost in longitudinal data is a significant methodological advance.
●
healthThis can improve the detection of rare clinical events in long-term patient histories.