Improving Supreme Court Forecasting Using Boosted Decision Trees

Aaron Kaufman, Peter Kraft, Maya Sen

Research output: Contribution to journalArticle

Abstract

Though used frequently in machine learning, boosted decision trees are largely unused in political science, despite many useful properties. We explain how to use one variant of boosted decision trees, AdaBoosted decision trees (ADTs), for social science predictions. We illustrate their use by examining a well-known political prediction problem, predicting U.S. Supreme Court rulings. We find that our ADT approach outperforms existing predictive models. We also provide two additional examples of the approach, one predicting the onset of civil wars and the other predicting county-level vote shares in U.S. presidential elections.

Original languageEnglish (US)
Pages (from-to)381-387
Number of pages7
JournalPolitical Analysis
DOIs
StateAccepted/In press - Jan 1 2019

Fingerprint

Supreme Court
predictive model
presidential election
civil war
political science
voter
social science
learning

Keywords

  • forecasting
  • Learning
  • statistical analysis of texts

ASJC Scopus subject areas

  • Sociology and Political Science
  • Political Science and International Relations

Cite this

Improving Supreme Court Forecasting Using Boosted Decision Trees. / Kaufman, Aaron; Kraft, Peter; Sen, Maya.

In: Political Analysis, 01.01.2019, p. 381-387.

Research output: Contribution to journalArticle

Kaufman, Aaron ; Kraft, Peter ; Sen, Maya. / Improving Supreme Court Forecasting Using Boosted Decision Trees. In: Political Analysis. 2019 ; pp. 381-387.
@article{977951d97c3a4b51818d042751ae510d,
title = "Improving Supreme Court Forecasting Using Boosted Decision Trees",
abstract = "Though used frequently in machine learning, boosted decision trees are largely unused in political science, despite many useful properties. We explain how to use one variant of boosted decision trees, AdaBoosted decision trees (ADTs), for social science predictions. We illustrate their use by examining a well-known political prediction problem, predicting U.S. Supreme Court rulings. We find that our ADT approach outperforms existing predictive models. We also provide two additional examples of the approach, one predicting the onset of civil wars and the other predicting county-level vote shares in U.S. presidential elections.",
keywords = "forecasting, Learning, statistical analysis of texts",
author = "Aaron Kaufman and Peter Kraft and Maya Sen",
year = "2019",
month = "1",
day = "1",
doi = "10.1017/pan.2018.59",
language = "English (US)",
pages = "381--387",
journal = "Political Analysis",
issn = "1047-1987",
publisher = "Oxford University Press",

}

TY - JOUR

T1 - Improving Supreme Court Forecasting Using Boosted Decision Trees

AU - Kaufman, Aaron

AU - Kraft, Peter

AU - Sen, Maya

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Though used frequently in machine learning, boosted decision trees are largely unused in political science, despite many useful properties. We explain how to use one variant of boosted decision trees, AdaBoosted decision trees (ADTs), for social science predictions. We illustrate their use by examining a well-known political prediction problem, predicting U.S. Supreme Court rulings. We find that our ADT approach outperforms existing predictive models. We also provide two additional examples of the approach, one predicting the onset of civil wars and the other predicting county-level vote shares in U.S. presidential elections.

AB - Though used frequently in machine learning, boosted decision trees are largely unused in political science, despite many useful properties. We explain how to use one variant of boosted decision trees, AdaBoosted decision trees (ADTs), for social science predictions. We illustrate their use by examining a well-known political prediction problem, predicting U.S. Supreme Court rulings. We find that our ADT approach outperforms existing predictive models. We also provide two additional examples of the approach, one predicting the onset of civil wars and the other predicting county-level vote shares in U.S. presidential elections.

KW - forecasting

KW - Learning

KW - statistical analysis of texts

UR - http://www.scopus.com/inward/record.url?scp=85066436426&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85066436426&partnerID=8YFLogxK

U2 - 10.1017/pan.2018.59

DO - 10.1017/pan.2018.59

M3 - Article

SP - 381

EP - 387

JO - Political Analysis

JF - Political Analysis

SN - 1047-1987

ER -