The deviance residual plots for these best fit models are given in Figure 10.
A residual analysis to detect outliers and influential data was performed using deviance residual plots.
Additionally, there was a covariate pattern with a standardized residual of -2.12 (deviance residual of -0.97).
The deviance residual plots revealed no obvious outliers in the data (data not shown).
There was one covariate pattern that had a standardized Pearson residual of 2.96 (deviance residual of 1.06).
Based on the deviance residual plots for both methods BYM and MLE, we found that there is no serious lack of fit in the model.
The average cross-validated deviance residual over V subsamples is Let γ k * be the value of that minimizes R cv of, k = 0,1,2,⋯.
The deviance residual provides a measure of how the survival of the proband compares to other probands with the same age and year of diagnosis.
In the application of the cure survival CART with exponential distribution, based on the average of 500 replications, the minimum cross-validated deviance residual was R cv (γ k *) = 1.753 (SE = 0.083), and the corresponding tree was quite large.
We first needed to classify the prognosis of the proband, and we did so based on the deviance residual from a multivariable proportional hazards (Cox) model fitted to the proband data adjusting for period and age of diagnosis.
