Dataset UCSB Breast cancer

Basic characteristics UCSB Breast cancer

2002

instances

58

bags

708

features

26

positive bags

between 21 and 40

instances per bag

32

negative bags

Results UCSB Breast cancer

AUC dataset 332
classifier error
AUC EER meanPrec avCost
APR - - - -
DiverseDensity (10.000000) 89.3 (6.0) 13.6 (7.5) 90.0 (3.8) 4.6 (2.1)
EMDD (10.000000) - - - -
MILBoost (100 rounds) 83.4 (2.2) 23.0 (2.4) 86.3 (1.9) 7.2 (0.8)
Citation 1NN, c=1 66.6 (4.5) 35.3 (5.3) 66.8 (2.1) 16.1 (1.2)
Citation 3NN, c=5 73.1 (1.3) 34.4 (4.9) 78.7 (1.6) 11.5 (0.7)
MI-SVM p=1.000000 88.9 (2.3) 16.2 (4.3) 90.5 (1.6) 4.9 (1.0)
MI-SVM r=10.000000 89.5 (2.8) 13.6 (3.8) 91.1 (2.0) 4.7 (1.2)
MILES p=1.000000 81.6 (4.5) 24.6 (4.9) 85.2 (4.0) 8.2 (1.7)
MILES r=10.000000 74.9 (2.4) 29.4 (5.0) 80.5 (1.9) 10.0 (0.9)
SVmil mahal.K 63.8 (26.9) 37.2 (23.1) 74.3 (17.2) 11.8 (7.1)
SVmil haussd.K 61.3 (21.4) 41.8 (16.3) 69.8 (12.5) 14.2 (4.4)
SVmil emd.K 67.4 (30.0) 36.9 (24.9) 77.1 (19.4) 11.0 (7.3)
SimpleMIL with Logistic2 52.1 (1.4) 47.9 (1.4) 45.8 (0.8) 24.4 (0.4)
mean-inst+Logistic2 90.9 (2.4) 14.0 (4.1) 92.9 (1.7) 4.2 (1.0)
extremes+Logistic2 85.6 (3.3) 22.7 (2.8) 88.7 (2.3) 6.7 (0.5)
cov-coef+Logistic2 92.2 (3.1) 14.4 (5.5) 93.3 (2.9) 4.0 (1.3)
Bag of Words (k=10)+Logistic2 50.8 (1.5) 49.2 (1.5) 45.6 (1.1) 24.6 (0.4)
Bag of Words (k=100)+Logistic2 61.5 (4.2) 40.7 (3.9) 62.6 (4.4) 18.0 (1.1)
PposteriorMIL 50.0 (0.0) 50.0 (0.0) 44.6 (0.0) 25.0 (0.0)
mean-inst+LIBSVM - - - -
extremes+LIBSVM - - - -
cov-coef+LIBSVM - - - -
Bag of Words (k=10)+LIBSVM - - - -
Bag of Words (k=100)+LIBSVM - - - -

Results are obtained using 10-fold stratified crossvalidation. The best results are indicated in bold, together with the results that are not significantly worse (based on a paired-difference t-test with a confidence level of 0.05). The standard deviations are shown between brackets.

Dissim.repr. AUC dataset 332
classifier error
AUC EER meanPrec avCost
minmin.K+Logistic2 81.6 (3.0) 23.7 (5.0) 85.0 (2.1) 8.1 (1.0)
summin.K+Logistic2 51.9 (11.2) 47.8 (5.6) 65.8 (7.0) 16.2 (2.2)
meanmin.K+Logistic2 83.1 (5.4) 22.3 (4.8) 87.3 (3.1) 7.0 (1.0)
meanmean.K+Logistic2 - - - -
mahal.K+Logistic2 - - - -
haussd.K+Logistic2 - - - -
emd.K+Logistic2 - - - -
linass.K+Logistic2 - - - -

Results are obtained using 10-fold stratified crossvalidation. The best results are indicated in bold, together with the results that are not significantly worse (based on a paired-difference t-test with a confidence level of 0.05). The standard deviations are shown between brackets.