moaftme: mixture of accelerated failure time experts


MOAFTME (Mixture of Accelerated Failure Time Model Experts) is an R package for the estimation of parameters of a survival analysis model consisting in a mixture of experts, each of which is an accelerated failure time (AFT) model. Estimation is performed witn an MCMC algorithm.

This package was written as a project for a course in Bayesian statistics. While the core is functional and relatively well optimized in terms of speed (many parts are in C), the current interface is rather simplistic.

I learned R as I went along. It's a nice language, it has some nice functional flavor to it and it's unbeatable for data analysis. The core of the library is pretty decent in my opinion, and it works just fine giving it 2D matrices. But I never quite figured out how to implement the traditional "R-like" interfaces with formulas and so on. Someday.