{"id":"arxiv/2609.10904","url":"https://arxiv.org/abs/2609.10904","tags":["stat.ME","math.ST"],"score":0,"title":"Agnostic Model-Assisted Estimation with Machine Learning for Survey Data","domain":"stat.ME","source":"arxiv","authors":["Ziming An","Mehdi Dagdoug","David Haziza","Yves Tillé"],"reviews":0,"subject":"stat.ME","summary":"Model-assisted estimation uses prediction rules to improve the efficiency of estimators of finite population parameters while retaining design-based inference. Although flexible prediction methods have been considered, existing theoretical results are largely method-specific. We develop a learner-agnostic framework that replaces separate analyses for individual learners with general conditions on the sampling design and prediction error. We connect design-aware and design-agnostic cross-fitting and characterize the sampling designs under which they yield conditional independence across…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-09-09","reactions":{}}