Skip to main content

Table 12 Relative contrast using Euclidean distance

From: D-Confidence: an active learning strategy to reduce label disclosure complexity in the presence of imbalanced class distributions

Dataset

Dim.

Instances

min.rc

max.rc

global.rc

Iris

4

150

4.477

63.985

69.852

Vowels

10

330

2.424

51.371

65.253

Poker

10

500

2.117

7.426

9.630

Cleveland

13

298

1.085

59.944

68.921

Satlog

36

500

1.682

23.760

26.258

R52

6019

1000

0.315

52.292

67.812

NG

10333

500

0.428

23.034

32.197