AUG 12, 2026 · PREPRINT
JAPE: Joint Anomaly Prediction and Intrinsic Explanation in Multivariate Time Series
arXiv
This is an unreviewed preprint presenting a novel machine-learning framework for anomaly prediction in time series; it reports experimental results on benchmarks but lacks peer review and clinical validation.
Reported
F1 improvement19.7%
AUC-PR improvement41.3%
MRR (explainability) gain26.6%