Nationality Classification Using Name Embeddings

Nov 6, 2016

Nationality identi€cation unlocks important demographic information, with many applications in biomedical and sociological research. Existing name-based nationality classi€ers use name substrings as features and are trained on small, unrepresentative sets of labeled names, typically extracted from Wikipedia. As a result, these methods achieve limited performance and cannot support €ne-grained classi€cation. We exploit the phenomena of homophily in communication patterns to learn name embeddings, a new representation that encodes gender, ethnicity, and nationality which is readily applicable to building classi€ers and other systems. Œrough our analysis of 57M contact lists from a major Internet company, we are able to design a €ne-grained nationality classi€er covering 39 groups representing over 90% of the world population. In an evaluation against other published systems over 13 common classes, our F1 score (0.795) is substantial beŠer than our closest competitor Ethnea (0.580). To the best of our knowledge, this is the most accurate, €ne-grained nationality classi€er available. As a social media application, we apply our classi€ers to the followers of major TwiŠer celebrities over six di‚erent domains. We demonstrate stark di‚erences in the ethnicities of the followers of Trump and Obama, and in the sports and entertainments favored by di‚erent groups. Finally, we identify an anomalous political €gure whose presumably inƒated following appears largely incapable of reading the language he posts in

  • ACM Conference on Information and Knowledge Management ([CIKM])
  • Conference/Workshop Paper