Predicting Censored Count Data with COM-Poisson Regression

20 Pages Posted: 6 Nov 2010 Last revised: 25 Jul 2011

Kimberly F. Sellers

Georgetown University - Department of Mathematics and Statistics

Galit Shmueli

Institute of Service Science, National Tsing Hua University, Taiwan

Date Written: October 29, 2010

Abstract

Censored count data are encountered in many applications, often due to a data collection mechanism that introduces censoring. A common example is questionnaires with question answers of the type 0,1,2,3. We consider the problem of predicting a censored output variable Y, given a set of complete predictors X. The common solution would be to use adaptations for Poisson or negative binomial regression models that account for the censoring. We study two alternatives that allow for both over- and under-dispersion: Conway-Maxwell-Poisson (COM-Poisson) regression, and generalized Poisson regression models, each with adaptations for censoring. We compare the predictive power of these models by applying them to a German panel dataset on fertility, where we introduce censoring of di erent levels into the outcome variable. We explore two additional variants: (1) using the mean versus the median of the predictive count distribution, and (2) ensembles of COM-Poisson models based on the parametric and non-parametric bootstrap.

Keywords: over-dispersion, under-dispersion, predictive distribution, mean versus median predictions, ensembles

Suggested Citation

Sellers, Kimberly F. and Shmueli, Galit, Predicting Censored Count Data with COM-Poisson Regression (October 29, 2010). Robert H. Smith School Research Paper No. RHS-06-129. Available at SSRN: https://ssrn.com/abstract=1702845 or http://dx.doi.org/10.2139/ssrn.1702845

Kimberly F. Sellers

Georgetown University - Department of Mathematics and Statistics ( email )

United States
202-687-8829 (Phone)

HOME PAGE: http://www9.georgetown.edu/faculty/kfs7

Galit Shmueli (Contact Author)

Institute of Service Science, National Tsing Hua University, Taiwan ( email )

Hsinchu, 30013
Taiwan

HOME PAGE: http://www.iss.nthu.edu.tw

Paper statistics

Downloads
259
Rank
96,291
Abstract Views
1,226