Summary
Survival data is different to typical regression data, and the incomplete observations pose a challenge. Since the data structure is completely different, we need specialized techniques to handle the incomplete observations and to that end, we introduced core survival concepts, such as hazard rate and survival function. We then introduced parametric lifetime models, which gives us a brief peek at how the lifetime distribution should look. We even fitted these lifetime distributions into the pbc dataset.
We also learned that the parametric setup might be very restrictive, and hence considered the nonparametric methods of the estimation of survival quantities. We also demonstrated the utility of the Nelson-Aalen estimator, the Kaplan-Meier survival function, and the log-rank test. The parametric hazards regression model was backed with the Cox proportional hazards regression model and applied to the pbc dataset. The logrank test can also help in the splitting criteria, and it has also...
