Skip to main content
. 2014 Oct 27;34(2):317–342. doi: 10.1002/sim.6343

Table 12.

Summary of the cross-validation results in the ER data set and the CHOP data set.

ER CHOP
nm+ pvmcv+ pvmcv HRmcv+ HRmcv nm+ pvmcv+ pvmcv HRmcv+ HRmcv
PRIM 80 (11) 0.10 (0.09) 0.86 (0.10) 0.59 (0.14) 1.50 (0.26) 82 (13) 0.012 (0.017) 0.10 (0.09) 0.44 (0.12) 0.66 (0.13)
Re-PRIM 86 (13) 0.30 (0.19) 0.71 (0.15) 0.81 (0.20) 1.22 (0.21) 87 (15) 0.009 (0.013) 0.14 (0.15) 0.44 (0.10) 0.70 (0.13)
AIM 148 (9) 0.23 (0.09) 0.93 (0.06) 0.77 (0.08) 2.35 (0.73) 136 (15) 0.043 (0.039) 0.21 (0.15) 0.60 (0.09) 0.71 (0.17)
PRIM (M) 102 (14) 0.12 (0.10) 0.88 (0.09) 0.67 (0.14) 1.63 (0.41) 108 (16) 0.010 (0.013) 0.17 (0.14) 0.49 (0.11) 0.71 (0.17)
Re-PRIM (M) 109 (11) 0.28 (0.16) 0.79 (0.16) 0.82 (0.13) 1.38 (0.34) 102 (13) 0.003 (0.004) 0.26 (0.17) 0.42 (0.09) 0.80 (0.16)

Note: Median absolute deviations are shown in braces. The letter M in braces indicates the approach of discovering disjunction of multiple conjunctive rule sets via covering.