Skip to main content

Table 3 The ranks of variable importance on hnscc dataset

From: A comparative study of forest methods for time-to-event data: variable selection and predictive performance

Characteristics Median ranks of VIM Frequency of VIM ranking top 38
RSF CIF MSR-RF RSF CIF MSR-RF
AFF3 9 5 8 0.81 0.99 0.85
PCDP1 49 36.5 52 0.43 0.54 0.35
ADH1B 12 23.5 39.5 0.87 0.8 0.48
RIMKLA 63 80.5 64.5 0.21 0.07 0.12
PCDHA12 58 64 56.5 0.24 0.21 0.14
CDHR5 48 11.5 32 0.43 0.94 0.56
PGC 4 19.5 37.5 0.94 0.87 0.51
NME5 48 43 45 0.31 0.43 0.43
XIST 36.5 27.5 50 0.51 0.66 0.33
B3GAT1 15 17.5 19 0.72 0.91 0.75
C1orf88 14 6 5 0.8 0.97 0.98
SELENBP1 17 25 15.5 0.73 0.77 0.75
ASRGL1 56.5 41.5 39 0.36 0.47 0.5
SGK2 52 15.5 45.5 0.36 0.86 0.44
PPP1R9A 49 81.5 66 0.29 0.05 0.21
KLF15 48 54.5 50 0.34 0.3 0.35
PNMA2 67 74 56.5 0.16 0.08 0.26
DNALI1 45 58 36 0.38 0.29 0.53
TMEM178 62.5 74 58.5 0.28 0.08 0.25
GAL3ST1 31.5 9 17 0.57 0.93 0.71
PLCXD3 42.5 64 55.5 0.47 0.13 0.19
RSPH1 44.5 24.5 48 0.44 0.7 0.4
METTL7B 31 17 25 0.58 0.88 0.62
DCDC2 63.5 64 61.5 0.17 0.19 0.16
DDAH1 64.5 56 44 0.24 0.26 0.46
RGL3 58.5 72 59.5 0.23 0.1 0.18
C16orf89 54.5 77 62 0.31 0.05 0.26
TCF21 61.5 66 51.5 0.24 0.16 0.33
C8orf46 62.5 70 66 0.21 0.11 0.32
CTSE 23 14 10 0.67 0.9 0.81
GJB1 60 77 64 0.18 0.06 0.22
BEX1 31 28 26.5 0.6 0.67 0.56
ABCC6P2 57 46.5 46 0.31 0.38 0.44
MYEF2 45.5 37 44 0.46 0.52 0.43
SLC4A4 63.5 62 61 0.18 0.19 0.3
CA8 64.5 68 61.5 0.23 0.13 0.19
CACNA1D 44 43 41 0.41 0.43 0.45
AQP4 62 77.5 58 0.2 0.06 0.22
median 48.5 44.75 47 0.36 0.40 0.42
range 4 ~ 67 5 ~ 81.5 5 ~ 66 0.16 ~ 0.94 0.05 ~ 0.99 0.12 ~ 0.98
IQR 32.75 ~ 61.12 23.75 ~ 67.5 36.38 ~ 58.38 0.24 ~ 0.56 0.13 ~ 0.79 0.25 ~ 0.53