wnpmle - Weighted NPMLE for Recurrent Events with a Competing Terminal Event

Provides regression modeling and prediction for the marginal mean of recurrent events in the presence of a competing terminal event based on the weighted nonparametric maximum likelihood estimator (wNPMLE) of Bellach and Kosorok (2026) <https://arxiv.org/abs/2605.25934>. Two large classes of transformation models are included: the Box-Cox transformation models and the logarithmic transformation models, extending the proportional means model of Ghosh and Lin (2002) <doi:10.17615/pt0g-y207> and the model of Zeng and Lin (2006) <doi:10.1093/biomet/93.3.627>. Parameter estimates are derived via automatic differentiation using the Template Model Builder (TMB) framework. Standard errors are computed via a sandwich variance estimator with correction for the estimated weights, following Bellach, Kosorok, Ruschendorf and Fine (2019) <doi:10.1080/01621459.2017.1401540>.