Skip to main content

Table 5 Average c-index values of the proposed model and random forest model using different amount of training data. The mean c-index was obtained by randomly selecting the training and test dataset 50 times

From: Investigating the relevance of major signaling pathways in cancer survival using a biologically meaningful deep learning model

Ratio of training data (%) Proposed model Random forest model
Mean c_index on training data Mean c_index on test data Mean c_index on training data Mean c_index on test data
GBM
50 0.6672 0.5869 0.7202 0.5383
60 0.6630 0.6033 0.7034 0.5404
70 0.6745 0.6029 0.6982 0.5450
80 0.6568 0.6085 0.6936 0.5493
90 0.6636 0.6392 0.6929 0.5612
SKCM
50 0.7959 0.6961 0.7975 0.5178
60 0.7576 0.6680 0.7879 0.5237
70 0.7209 0.6950 0.7751 0.5277
80 0.7685 0.6717 0.7653 0.5255
90 0.7246 0.6643 0.7541 0.5208
BRCA     
50 0.6235 0.5435 0.8356 0.5217
60 0.6723 0.5768 0.8298 0.5262
70 0.6942 0.5627 0.8222 0.5154
80 0.7038 0.5835 0.8075 0.5026
90 0.7076 0.6157 0.8069 0.5036
LUAD
50 0.7043 0.5857 0.8486 0.5641
60 0.7345 0.6708 0.8294 0.5702
70 0.6954 0.6363 0.8187 0.5704
80 0.7560 0.7249 0.8088 0.5809
90 0.7394 0.7419 0.8020 0.5822