Multi-Label Prediction for Political Text-as-Data

Authors

Aaron Erlich, McGill University
Stefano G. Dantas, McGill University
Benjamin E. Bagozzi, University of Delaware
Daniel Berliner, London School of Economics and Political Science
Brian Palmer-Rubin, Marquette UniversityFollow

Document Type

Article

Publication Date

10-2022

Publisher

Cambridge University Press

Source Publication

Political Analysis

Source ISSN

1047-1987

Original Item ID

DOI: 10.1017/pan.2021.15

Abstract

Political scientists increasingly use supervised machine learning to code multiple relevant labels from a single set of texts. The current “best practice” of individually applying supervised machine learning to each label ignores information on inter-label association(s), and is likely to under-perform as a result. We introduce multi-label prediction as a solution to this problem. After reviewing the multi-label prediction framework, we apply it to code multiple features of (i) access to information requests made to the Mexican government and (ii) country-year human rights reports. We find that multi-label prediction outperforms standard supervised learning approaches, even in instances where the correlations among one’s multiple labels are low.

Comments

Accepted version. Political Analysis, Vol. 30, No. 4 (October 2022): 463-480. DOI.

Recommended Citation

Erlich, Aaron; Dantas, Stefano G.; Bagozzi, Benjamin E.; Berliner, Daniel; and Palmer-Rubin, Brian, "Multi-Label Prediction for Political Text-as-Data" (2022). Political Science Faculty Research and Publications. 133.
https://epublications.marquette.edu/polisci_fac/133