Splitting It Up: The spduration Split-Population Duration Regression Package for Time-Varying Covariates

Abstract

We present an implementation of split-population duration regression in the spduration (Beger et al., 2017) package for R that allows for time-varying covariates. The statistical model accounts for units that are immune to a certain outcome and are not part of the duration process the researcher is primarily interested in. We provide insights for when immune units exist, that can significantly increase the predictive performance compared to standard duration models. The package includes estimation and several post-estimation methods for split-population Weibull and log-logistic models. We provide an empirical application to data on military coups.

Reassessing the Role of Theory and Machine Learning in Forecasting Civil Conflict

“…For example, both Ward, Greenhill, and Bakke (2010) and Hill and Jones (2014) examine a variety of variables that are considered “theoretically” important for explaining civil war onset and state repression, respectively, and find that there is substantial variation in how useful they are for actually predicting those outcomes. It is also possible that building accurate predictive models could uncover factors overlooked in the literature, for example, the strong association between infant mortality and a broad variety of conflict outcomes like political instability (Goldstone et al 2010), irregular leadership changes (Beger et al 2017), and coups (Beger and Ward 2020).…”

“…The implementation error aside, this split-population analog model is quite odd and does not replicate the idea behind split-population modeling (Chiba, Metternich, and Ward 2015; Beger et al 2017). Although the two random forest models are trained on separate data (in our updated, fixed replication), the process of combining them actually creates a new, larger random forest using both component model’s underlying decision trees.…”

BayesMFSurv: An R Package to Estimate Bayesian Split-Population Survival Models With (and Without) Misclassified Failure Events

“…Numerous R packages offer functionalities to estimate conventional parametric and semiparametric survival models via maximum likelihood estimation (MLE) or Bayesian MCMC methods (Diez, 2013;Therneau, 2019;Wang, Chen, Wang, & Yan, 2019;Zhou, Hanson, & Zhang, 2020). Other R packages focus on estimation of parametric or semi-parametric cure survival models using MLE (Amdahl, 2019;Beger, Hill, Metternich, Minhas, & Ward, 2017;Cai, Zou, Peng, & Zhang, 2012;Han, Zhang, & Shao, 2017). To our knowledge, there is no R package that fits parametric mixture cure models, including the MF survival model, via Bayesian MCMC (e.g., slice sampling) methods that offer a powerful yet flexible tool for estimating such models.…”

A Bayesian Split Population Survival Model for Duration Data With Misclassified Failure Events

“…As such, our model shares similarities with the cure survival model, which has been previously used in Political Science to model competing processes of democratic survival (Svolik 2008), or to accommodate heterogeneous mixtures of “at risk” and “not at risk” countries in global analyses of irregular leadership changes (Beger, Dorff, and Ward 2014, 2015; Beger et al. 2017). However, in contrast to the cure model—which only allows one to model heterogeneous mixtures among observations that have not failed—our proposed model allows one to specifically account for heterogeneous mixtures of failure cases.…”