Online Survival Analysis: A Bandit Approach under Cox PH Model
Researchers have developed a novel bandit approach for online survival analysis, integrating the Cox proportional hazards model into sequential decision-making. This method addresses challenges like staggered entry, delayed feedback, and right censoring, adapting existing bandit algorithms to optimize treatments as new data emerges. Simulations and experiments with SEER cancer data indicate the approach effectively learns near-optimal treatment policies. AI
IMPACT Introduces a new framework for sequential decision-making in survival analysis, potentially improving treatment optimization in medical settings.