Variable Selection Methods for Right-Censored Time-to-Event Data with High-Dimensional Covariates

Algorithm 3

SSG algorithm.

Step  1. for = 1 to do

Step  2.  Split the data into equally sized subsets

Step  3.  Compose the dataset for each subset

Step  4.  Calculate over the for each subset

Step  5.  Compare and with log-rank test and save the test score for each subset one by one

Step  6.  Select a subset with the highest test score

Step  7.   Save the test score for variables in the selected subset as

Step  8. end for

Step  9. Assume = as the reverse variable efficiency vector where initially each array

   as the cumulative contribution score corresponding to a variable is zero

Step  10. for = 1 to do

Step  11.   for = 1 to do

Step  12.   Add the value of to the cumulative contribution score of the variable

     based on its identification number =

     end for

Step  13. end for

Step  14. Return = as the variable inefficiency vector