Skip to main content

Table 5 TCGA dataset validation

From: binomialRF: interpretable combinatoric efficiency of random forests to identify biomarker interactions

Model Time Test error Model size
5A. Breast cancer
 binomialRF 83 (11) 0 (0) 27 (4)
 RFE 100 (13) 0 (0) 692 (23)
 Perm 112 (16) 0 (0) 1092 (39)
 Vita 493 (88) 0 (0) 19,933 (10)
 Boruta 1667 (617) 0 (0) 92 (3)
5B. Kidney cancer
 binomialRF 51 (10) 0 (0) 48 (3)
 RFE 67 (10) 0 (0) 592 (55)
 Perm 73 (12) 0 (0) 867 (55)
 Vita 315 (72) 0 (0) 19,760 (41)
 Boruta 987 (363) 0 (0) 24 (2)
  1. The algorithms in Table 1 were tested and compared using the TCGA breast cancer and kidney datasets, reporting the mean (and standard deviation in parentheses). Half of the methods were not included as they encountered computation or memory limitations in running the TCGA datasets