A0244
Title: A complete guide to small area learning
Authors: Katarzyna Reluga - University of Bristol (United Kingdom) [presenting]
Abstract: Sample surveys are widely recognized as high-quality and cost-effective sources of information for obtaining estimates of target parameters at the population and subpopulation levels. If the sample size of the subpopulation is small (or even zero in some areas), researchers encounter the small area estimation (SAE) dilemma. SAE techniques have been developed to provide official statistics by leveraging survey samples and parametric statistical modelling. We introduce a general framework for small area learning (SAL). SAL encompasses machine learning to obtain estimates of subpopulation-level parameters by pooling information from other subpopulations, which is the main principle of classical SAE. In addition to presenting a complete methodological setup for inference and prediction, we provide a practical application of SAL in measuring poverty.
