Primal--Dual Alternating Neural Learning for Timely Classification with Performance Guarantees

The paper introduces a new method for timely risk classification in clinical monitoring, framing the problem as a multi‑objective optimization that balances early classification, sensitivity, specificity, and monitoring cost. It derives an optimal decision rule via a value recursion and estimates it from data using a recurrent neural network combined with a primal–dual updating scheme to enforce performance constraints. Experiments, including a case study on continuous glucose monitoring for hypoglycemia prediction, show that the approach produces accurate, timely decision rules that meet the specified operating characteristics.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Statistics ML.