
method	CA	AP	Brier	IS
bayes	0.903	0.902	0.176	 0.758
tree	0.825	0.824	0.326	 0.599
majrty	0.614	0.526	0.474	-0.000

method	CA	AP	Brier	IS
bayes	0.903+-0.008	0.902+-0.008	0.176+-0.016	 0.758+-0.017
tree	0.825+-0.016	0.824+-0.016	0.326+-0.033	 0.599+-0.034
majrty	0.614+-0.003	0.526+-0.001	0.474+-0.001	-0.000+-0.000

Confusion matrix for naive Bayes:
TP: 238, FP: 13, FN: 29.0, TN: 155

Confusion matrix for naive Bayes:
TP: 240, FP: 18, FN: 27.0, TN: 150

Confusion matrix for naive Bayes for 'van':
TP: 192, FP: 152, FN: 7.0, TN: 495

Confusion matrix for naive Bayes for 'opel':
TP: 80, FP: 71, FN: 132.0, TN: 563

	bus	opel	saab	van
bus	156	25	17	20
opel	6	80	61	65
saab	7	46	97	67
van	4	0	3	192

Sensitivity and specificity for 'voting'
method	sens	spec
bayes	0.891	0.923
tree	0.801	0.863
majrty	1.000	0.000

Sensitivity and specificity for 'vehicle=van'
method	sens	spec
bayes	0.965	0.765
tree	0.794	0.969
majrty	0.000	1.000

AUC (voting)
     bayes: 0.974
      tree: 0.926
    majrty: 0.500

AUC for vehicle using weighted single-out method
bayes	tree	majority
0.841	0.795	0.500

AUC for vehicle, using different methods
                           	bayes	tree	majority
       by pairs, weighted: 	0.858	0.869	0.500
                 by pairs: 	0.859	0.870	0.500
    one vs. all, weighted: 	0.841	0.795	0.500
              one vs. all: 	0.841	0.795	0.500

AUC for detecting class 'van' in 'vehicle'
0.924	0.881	0.500

AUCs for detecting various classes in 'vehicle'
bus (218.000) vs others: 	0.954	0.943	0.500
opel (212.000) vs others: 	0.749	0.685	0.500
saab (217.000) vs others: 	0.739	0.672	0.500
van (199.000) vs others: 	0.924	0.881	0.500

	bus	opel	saab
opel	0.922
saab	0.927	0.561
van	0.991	0.898	0.857

AUCs for detecting various pairs of classes in 'vehicle'
opel vs bus: 	0.922	0.949	0.500
saab vs bus: 	0.927	0.941	0.500
saab vs opel: 	0.561	0.578	0.500
van vs bus: 	0.991	0.977	0.500
van vs opel: 	0.898	0.902	0.500
van vs saab: 	0.857	0.872	0.500

AUC and SE for voting
bayes: 0.982+-0.008
tree: 0.888+-0.025
majrty: 0.500+-0.045

Difference between naive Bayes and tree: 0.065+-0.066

ROC (first 20 points) for bayes on 'voting'
1.000	1.000
0.970	1.000
0.940	1.000
0.910	1.000
0.896	1.000
0.881	1.000
0.836	1.000
0.821	1.000
0.806	1.000
0.761	1.000
0.746	1.000
0.731	1.000
0.701	1.000
0.687	1.000
0.672	1.000
0.627	1.000
0.612	1.000
0.597	1.000
0.582	1.000
0.567	1.000
