ACS Income for Domain Generalization

This product uses the Census Bureau Data API but is not endorsed or certified by the Census Bureau. This dataset is a preprocessed version of the American Community Survey (ACS) 2018 Public Use Microdata Sample (PUMS), provided by the U.S. Census Bureau. It is designed for machine learning research on domain generalization. The task is to predict individual income based on demographic and socioeconomic features, with age group used to define domain splits. Preprocessing was adapted from the Whyshift repository (https://github.com/namkoong-lab/whyshift), which builds on the Folktables project (https://github.com/socialfoundations/folktables). Folktables provides tools to extract and structure ACS PUMS data. The data remains governed by the U.S. Census Bureau terms of service, available at: https://www.census.gov/data/developers/about/terms-of-service.html.