Abstract:
Raking is a critical tool for adjusting inputs to match known totals—arising naturally in calibrating survey weights to census data and reconciling estimates in complex workflows. We review the underlying optimization problem, casting raking as minimizing entropic distance subject to linear constraints, and leverage this perspective to develop an efficient and versatile computational framework for modern applications.
We discuss practical extensions, including (1) differential weighting to account for varying input uncertainty, (2) raking bounded quantities using logistic distances, (3) raking inputs with incompatible margins, and (4) handling high dimensional applications. We illustrate using synthetics and show how the new raking package applies to a large-scale LSAE application recently presented at CSSS. The methods are implemented in a pip-installable python package with an intuitive API (and easy access through R).
Dr. Aleksandr Aravkin received his PhD in Mathematics (Optimization) and master’s degree in statistics from the University of Washington in 2010. After some time at IBM TJ Watson Research center, he returned to the UW Applied Mathematics department in 2015, where he is currently an Associate Professor, and adjunct with Mathematics, Statistics, Computer Science, and Health Metrics Sciences. Since 2019, Dr. Aravkin has also held the title of Director of Mathematical Sciences at the Institute for Health Metrics and Evaluation (IHME), developing new methods for emerging problems in global health.
