Cox proportional hazards model with Weibull base hazard rate

This function implements a Cox PH model with a Weibull base hazard rate (also called a Weibull proportional hazards model).
The standard Cox model assumes (usually implicitly) Breslow's non-parametric baseline hazard estimator. This is ill suited to predicting the event time for new individuals. For this purpose the Weibull-Cox model can provide predictions with error bars (given by the standard deviation) along with the usual regression coefficients, (smooth) survival functions and (smooth) hazard rates.
Functions are included for predicting the event time for an individual once a model has been trained.
Get started by running example.m.
Please email me with comments or requests.