Skip to main content
. Author manuscript; available in PMC: 2015 Nov 24.
Published in final edited form as: Proc SIAM Int Conf Data Min. 2015;2015:208–216. doi: 10.1137/1.9781611974010.24

Table 1.

Experimental Results on Simulated and Real datasets

(a) Non-Linear XOR configuration results
LR ACP IsoReg Platt Hist SBB ABB
AUC 0.497 0.950 0.704 0.497 0.931 0.914 0.941
Acc 0.510 0.887 0.690 0.510 0.855 0.887 0.888
RMSE 0.500 0.286 0.447 0.500 0.307 0.307 0.295
MCE 0.521 0.090 0.642 0.521 0.152 0.268 0.083
ECE 0.190 0.056 0.173 0.190 0.072 0.104 0.062
(b) Non-Linear Circular configuration results
LR ACP IsoReg Platt Hist SBB ABB
AUC 0.489 0.852 0.635 0.489 0.827 0.816 0.838
Acc 0.500 0.780 0.655 0.500 0.795 0.790 0.773
RMSE 0.501 0.387 0.459 0.501 0.394 0.393 0.390
MCE 0.540 0.172 0.608 0.539 0.121 0.790 0.146
ECE 0.171 0.098 0.186 0.171 0.074 0.138 0.091
(c) Adult Naïve Bayes
NB IsoReg Platt Hist SBB ABB
AUC 0.879 0.876 0.879 0.877 0.849 0.879
Acc 0.803 0.822 0.840 0.818 0.838 0.835
RMSE 0.352 0.343 0.343 0.341 0.345 0.343
MCE 0.223 0.302 0.092 0.236 0.373 0.136
ECE 0.081 0.075 0.071 0.078 0.114 0.062
(d) Adult Linear SVM
SVM IsoReg Platt Hist SBB ABB
AUC 0.864 0.856 0.864 0.864 0.821 0.864
Acc 0.248 0.805 0.748 0.815 0.803 0.805
RMSE 0.587 0.360 0.434 0.355 0.362 0.357
MCE 0.644 0.194 0.506 0.144 0.396 0.110
ECE 0.205 0.085 0.150 0.077 0.108 0.061
(e) Adult Logistic Regression
LR ACP IsoReg Platt Hist SBB ABB
AUC 0.730 0.727 0.732 0.730 0.743 0.699 0.731
Acc 0.755 0.783 0.753 0.755 0.753 0.762 0.762
RMSE 0.403 0.402 0.403 0.405 0.400 0.401 0.401
MCE 0.126 0.182 0.491 0.127 0.274 0.649 0.126
ECE 0.075 0.071 0.118 0.079 0.092 0.169 0.076
(f) SPECT Naïve Bayes
NB IsoReg Platt Hist SBB ABB
AUC 0.836 0.815 0.836 0.832 0.733 0.835
Acc 0.759 0.845 0.770 0.824 0.845 0.845
RMSE 0.435 0.366 0.378 0.379 0.368 0.374
MCE 0.719 0.608 0.563 0.712 0.347 0.557
ECE 0.150 0.141 0.148 0.145 0.149 0.157
(g) SPECT SVM Quadratic kernel
SVM IsoReg Platt Hist SBB ABB
AUC 0.816 0.786 0.816 0.766 0.746 0.810
Acc 0.257 0.834 0.684 0.845 0.813 0.813
RMSE 0.617 0.442 0.460 0.463 0.398 0.386
MCE 0.705 0.647 0.754 0.934 0.907 0.769
ECE 0.235 0.148 0.162 0.180 0.128 0.131
(h) SPECT Logistic Regression
LR ACP IsoReg Platt Hist SBB ABB
AUC 0.744 0.742 0.733 0.744 0.738 0.733 0.741
Acc 0.658 0.561 0.626 0.668 0.620 0.620 0.626
RMSE 0.546 0.562 0.558 0.524 0.565 0.507 0.496
MCE 0.947 1.000 1.000 0.884 0.997 0.813 0.812
ECE 0.181 0.187 0.177 0.180 0.183 0.171 0.173
(i) CAP Naïve Bayes
NB IsoReg Platt Hist SBB ABB
AUC 0.848 0.845 0.848 0.831 0.775 0.838
Acc 0.730 0.865 0.847 0.853 0.832 0.865
RMSE 0.504 0.292 0.324 0.307 0.315 0.304
MCE 0.798 0.188 0.303 0.087 0.150 0.128
ECE 0.161 0.071 0.097 0.056 0.067 0.067
(j) CAP Linear SVM
SVM IsoReg Platt Hist SBB ABB
AUC 0.858 0.858 0.858 0.847 0.813 0.863
Acc 0.907 0.900 0.882 0.887 0.902 0.908
RMSE 0.329 0.277 0.294 0.287 0.285 0.274
MCE 0.273 0.114 0.206 0.110 0.240 0.121
ECE 0.132 0.058 0.093 0.057 0.083 0.050
(k) CAP Logistic Regression
LR ACP IsoReg Platt Hist SBB ABB
AUC 0.920 0.910 0.917 0.920 0.901 0.856 0.921
Acc 0.925 0.932 0.935 0.928 0.897 0.935 0.932
RMSE 0.240 0.240 0.234 0.242 0.259 0.240 0.240
MCE 0.199 0.122 0.286 0.154 0.279 0.391 0.168
ECE 0.066 0.062 0.078 0.082 0.079 0.103 0.069