Dataset Corel Dinosaurs

Basic characteristics Corel Dinosaurs

7947

instances

2000

bags

9

features

100

positive bags

between 2 and 13

instances per bag

1900

negative bags

Results Corel Dinosaurs

AUC dataset 114
classifier error
AUC EER meanPrec avCost
APR 71.5 (0.1) 28.5 (0.1) 8.5 (0.0) 18.2 (0.0)
DiverseDensity (10.000000) 87.9 (0.1) 20.4 (0.8) 52.5 (0.5) 10.2 (0.2)
EMDD (10.000000) 99.8 (0.0) 0.5 (0.0) 97.4 (0.0) 0.4 (0.0)
MILBoost (100 rounds) 98.5 (0.4) 4.3 (1.2) 94.9 (0.4) 1.8 (0.2)
Citation 1NN, c=1 97.2 (0.3) 2.8 (0.3) 82.4 (2.1) 2.5 (0.3)
Citation 3NN, c=5 99.6 (0.1) 0.5 (0.0) 92.6 (1.1) 0.5 (0.0)
MI-SVM p=1.000000 91.8 (2.1) 8.5 (1.9) 75.5 (4.7) 4.8 (1.0)
MI-SVM r=10.000000 99.8 (0.0) 0.3 (0.1) 95.8 (0.3) 0.3 (0.1)
MILES p=1.000000 99.9 (0.0) 0.2 (0.1) 97.8 (0.7) 0.2 (0.0)
MILES r=10.000000 99.9 (0.0) 0.4 (0.1) 97.7 (0.3) 0.4 (0.1)
SVmil mahal.K 36.4 (1.9) 60.2 (1.7) 6.2 (0.7) 23.8 (0.3)
SVmil haussd.K 93.5 (5.2) 11.9 (5.0) 70.0 (3.8) 4.7 (1.3)
SVmil emd.K 93.6 (6.7) 9.6 (6.5) 73.8 (2.9) 3.3 (1.6)
SimpleMIL with Logistic2 93.4 (0.1) 9.5 (0.4) 38.6 (0.5) 4.8 (0.1)
mean-inst+Logistic2 99.4 (0.0) 1.0 (0.3) 90.2 (0.2) 0.8 (0.1)
extremes+Logistic2 98.8 (0.5) 1.4 (0.5) 92.0 (1.5) 1.2 (0.5)
cov-coef+Logistic2 95.3 (1.4) 5.6 (1.3) 77.5 (5.6) 4.1 (1.2)
Bag of Words (k=10)+Logistic2 93.0 (1.5) 11.5 (1.9) 43.5 (6.3) 6.5 (0.8)
Bag of Words (k=100)+Logistic2 99.3 (0.1) 2.4 (0.6) 92.0 (2.2) 1.1 (0.1)
PposteriorMIL 97.1 (0.8) 6.8 (1.5) 69.4 (3.9) 3.5 (0.5)
Dissim.repr. AUC dataset 114
classifier error
AUC EER meanPrec avCost
minmin.K+Logistic2 99.8 (0.0) 0.4 (0.1) 96.6 (1.1) 0.3 (0.1)
summin.K+Logistic2 99.6 (0.4) 0.5 (0.4) 95.1 (0.5) 0.4 (0.3)
meanmin.K+Logistic2 99.9 (0.0) 0.3 (0.1) 95.9 (1.3) 0.2 (0.0)
meanmean.K+Logistic2 99.5 (0.0) 0.8 (0.1) 91.9 (1.1) 0.8 (0.1)
mahal.K+Logistic2 - - - -
haussd.K+Logistic2 - - - -
emd.K+Logistic2 - - - -
linass.K+Logistic2 - - - -