Comparing Random Forest with Logistic Regression for Predicting Class-Imbalanced Civil War Onset Data. (4th January 2017)

• Record Type: Journal Article
• Authors: Muchlinski, David; Siroky, David; He, Jingrui; Kocher, Matthew
• Is Part Of: Political analysis. Volume 24:Number 1(2016:Spring)
• Page Start: 87
• Page End: 103
• Publication Date: 2017-01-04
• Subjects: Political science -- Methodology -- Periodicals; Electronic journals; 320.011
• DOI: 10.1093/pan/mpv024
• Languages: English
• ISSNs: 1047-1987

Abstract

Abstract : The most commonly used statistical models of civil war onset fail to correctly predict most occurrences of this rare event in out-of-sample data. Statistical methods for the analysis of binary data, such as logistic regression, even in their rare event and regularized forms, perform poorly at prediction. We compare the performance of Random Forests with three versions of logistic regression (classic logistic regression, Firth rare events logistic regression, and L 1 -regularized logistic regression), and find that the algorithmic approach provides significantly more accurate predictions of civil war onset in out-of-sample data than any of the logistic regression models. The article discusses these results and the ways in which algorithmic statistical methods like Random Forests can be useful to more accurately predict rare events in conflict data.