A Comparison of Rule Induction Using Feature Selection and the LEM2 Algorithm - Feature Selection for Data and Pattern Recognition - page 168

Table 8.8 Results of experiments—rule set complexity
Data set

LEM1 LEM2

Number of
Rules Conditions Rules Conditions

Australian Credit Approval 317 1,380 115 559

Breast Cancer—Slovenia 160 518 92 319

Breast Cancer—Wisconsin 255 638 164 421

Bupa Liver Disorders 234 684 164 487

Glass 111 334 82 262

Hepatitis 47 191 21 88

Image segmentation 71 273 38 144

Iris 18 43 13 30

Lymphography 77 217 26 74

Pima 444 1,547 287 1,025

Postoperative patients 50 175 28 98

Soybean 101 411 42 139

Primary Tumor 171 829 125 638

Wine Recognition 32 111 16 52

Table 8.9 Results of
experiments—run time, in
milliseconds

Data set LEM1 LEM2

Australian Credit Approval 457 273

Breast Cancer—Slovenia 38 54

Breast Cancer—Wisconsin 146 126

Bupa Liver Disorders 30 69

Glass 31 30

Hepatitis 33 26

Image segmentation 65 39

Iris 3 1

Lymphography 36 27

Pima 257 407

Postoperative patients 7 5

Soybean 415 158

Primary Tumor 341 134

Wine Recognition 25 9