The authors compare models that use the letters in a name to estimate recorded race and ethnicity. Their best model classifies 85% of held-out cases correctly using full names. With surnames alone, the best model classifies 81% correctly.
Cite: Rajashekar Chintalapati, Suriyan Laohaprapanon, and Gaurav Sood (2023). "Predicting Race and Ethnicity From the Sequence of Characters in a Name." arXiv:1805.02109, first posted in 2018 and revised in 2023. https://doi.org/10.48550/arXiv.1805.02109
BibTeX:
@misc{sood2023predicting,
author = {Rajashekar Chintalapati and Suriyan Laohaprapanon and Gaurav Sood},
title = {{Predicting Race and Ethnicity From the Sequence of Characters in a Name}},
year = {2023},
doi = {10.48550/arXiv.1805.02109},
url = {https://doi.org/10.48550/arXiv.1805.02109}
}