{"id":"arxiv/2606.17777","url":"https://arxiv.org/abs/2606.17777","tags":["stat.ME","math.ST","stat.ML"],"score":0,"title":"On Response-Adaptive Targeting Strategies for Multi-Treatment Experiments","domain":"stat.ME","source":"arxiv","authors":["Redouane Yagouti","Rémy Degenne","Emilie Kaufmann"],"reviews":0,"subject":"stat.ME","summary":"Response-adaptive randomization (RAR) in clinical trials aims to improve ethical and statistical efficiency by dynamically allocating patients to treatments based on observed outcomes. While RAR based on a target optimal allocation have been extensively studied for two-arms settings, their extension to multi-treatment experiments ($K \\geq 2$) remains theoretically fragmented, with most existing methods focusing on specific algorithms or restricted target allocations. In this paper, we introduce a unified framework for response-adaptive targeting, the $α$-Rebalancing Targeting Strategies…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-06-16","reactions":{}}