Table 8.

Test set area under the ROC curve (AUC) using normalized surprisal for three types of feature prediction model, and for the combined normalized surprisal of all three of them. The best AUC scores are shown in bold

Data set	Linear Kernel SVM	RBF Kernel SVM	Decision tree	Combined
abalone	0.50	0.51	0.43	0.48
acute	0.99	1.00	0.93	1.00
adult	0.64	0.64	0.53	0.61
annealing	0.73	0.79	0.84	0.82
arrhythmia	0.77	0.77	0.78	0.78
audiology	0.81	0.79	0.78	0.80
balance-scale	0.94	0.96	0.94	0.97
blood-transfusion	0.56	0.60	0.56	0.59
breast-cancer-wisconsin	0.94	0.96	0.96	0.96
car	0.95	0.91	0.96	0.97
chess	0.89	0.91	0.92	0.93
cmc	0.41	0.42	0.42	0.41
connect-4	0.51	0.54	0.75	0.65
credit-screening	0.83	0.84	0.84	0.85
cylinder-bands	0.62	0.75	0.63	0.69
dermatology	0.98	1.00	0.99	1.00
echocardiogram	0.64	0.68	0.65	0.67
ecoli	0.96	0.96	0.96	0.97
glass	0.61	0.65	0.65	0.65
haberman	0.67	0.66	0.66	0.67
hayes-roth	0.92	0.91	0.88	0.92
hepatitis	0.80	0.80	0.84	0.83
horse-colic	0.78	0.79	0.80	0.82
image	0.97	0.96	0.98	0.98
internet_ads	0.96	0.89	0.95	0.94
ionosphere	0.96	0.97	0.96	0.97
iris	0.99	1.00	0.99	1.00
letter-recognition	0.99	1.00	0.99	1.00
libras	0.89	0.88	0.89	0.89
magic	0.80	0.74	0.86	0.83
mammographic-masses	0.72	0.74	0.71	0.73
mushroom	1.00	1.00	1.00	1.00
nursery	0.99	1.00	1.00	1.00
ozone	0.37	0.37	0.34	0.34
page-blocks	0.88	0.83	0.90	0.89
parkinsons	0.54	0.67	0.65	0.64
pima-indians-diabetes	0.73	0.74	0.72	0.75
poker	0.56	0.57	0.53	0.56
secom	0.54	0.56	0.61	0.57
spambase	0.79	0.85	0.84	0.84
statlog	0.61	0.60	0.62	0.63
tae	0.60	0.48	0.49	0.55
tic-tac-toe	0.96	0.97	0.78	0.99
voting-records	0.93	0.94	0.91	0.95
wine	0.95	0.94	0.91	0.96
yeast	0.72	0.72	0.71	0.72
zoo	0.99	0.99	1.00	1.00
Number of data sets with the maximum AUC score	3	8	6	11