Model-Assisted Estimation Through Random Forests in Finite Population Sampling

In surveys, the interest lies in estimating finite population parameters such as totals and means. most some auxiliary information is available at estimation stage. This may be incorporated procedures to increase their precision. this article, we use random forests estimate functional relationship between survey variable variables. recent years, have become attractive National Statistical Offices now access a variety of data sources, potentially exhibiting large number observations on We establish theoretical properties model-assisted based derive corresponding variance estimators. A model-calibration procedure for handling multiple variables also discussed. The results simulation study suggest that proposed point perform well term bias, efficiency, coverage normal-based confidence intervals, wide settings. Finally, apply methods using radio audiences collected by M\'ediam\'etrie, French audience company.