Skip to main content
. 2020 Jul 13;36(Suppl 1):i389–i398. doi: 10.1093/bioinformatics/btaa462

Fig. 2.

Fig. 2.

The box plot for the performance of the following survival prediction models on 10 cancer types: Cox-ridge, Cox-LASSO, Cox-nnet and our model VAECox. We randomly split the data into training (80%) and test sets (20%). We repeated this process 10 times and obtained 10 C-index scores. The white triangle of each box denotes the average of 10 C-index scores. The optimal hyperparameters were selected by fivefold cross validation on the training set