Chapter 10. Ensembling Survival Models
The primary biliary cirrhosis data was introduced in first two chapters using the jackknife method. Observations in clinical trials are usually subject to censoring, and the jackknife method helps to complete incomplete observations through the idea of pseudo values. Since pseudo values are likely to be dependent on each other, the generalized estimating equation framework made it possible to estimate the impact of twelve covariates at the time of interest. The idea of pseudo values and the generalized estimating equation framework makes it easy for practitioners to interpret the results. However, this method might not be useful if the number of censored observations is exceptionally high. Furthermore, it is also preferable to have statistical methods that preserve the incompleteness of the observations and yet make good use of them. The general (linear) regression framework with time as the dependent variable and the error term following...
