Skip to main content

Table 7 The obtained results for MMCC from boosting algorithms using decision tree as a base learner

From: Boosting methods for multi-class imbalanced data classification: an experimental review

Dataset Algorithm
AdaBoost.MH SAMME CatBoost LogitBoost GradientBoost XGBoost MEBoost SMOTEBoost RUSBoost LightGBM AdaC1 AdaC2 AdaC3 AdaCost
Wine 0.8134 0.8318 0.9546 0.9155 0.9823 0.9343 0.9664 0.9013 0.923 0.9382 0.865 0.8388 0.9176 0.8696
± 0.0254 ± 0.0236 ± 0.0086 ± 0.0188 ± 0.0050 ± 0.0237 ± 0.0155 ± 0.0325 ± 0.0178 ± 0.0280 ± 0.0320 ± 0.0327 ± 0.0254 ± 0.0262
Hayes-Roth 0.6109 0.5911 0.5229 0.6026 0.611 0.5967 0.5916 0.5876 0.5401 0.0649 0.6027 0.622 0.6325 0.6157
± 0.0274 ± 0.0212 ± 0.0267 ± 0.0228 ± 0.0222 ± 0.0302 ± 0.0374 ± 0.0300 ± 0.0554 ± 0.0682 ± 0.0255 ± 0.0199 ± 0.0288 ± 0.0325
Contraceptive 0.0562 − 0.0104 0.0615 − 0.0181 − 0.0111 0.073 0.0078 − 0.0395 − 0.1068 0.0018 0.0894 − 0.3029 − 0.1027 0.1025
± 0.0098 ± 0.0142 ± 0.0089 ± 0.0119 ± 0.0199 ± 0.0143 ± 0.0099 ± 0.0239 ± 0.0074 ± 0.0086 ± 0.0111 ± 0.0044 ± 0.0237 ± 0.0185
Pen-based 0.8073 0.918 0.9252 0.9125 0.9049 0.8684 0.946 0.9026 0.92 0.9076 0.8119 0.8121 0.8105 0.8111
± 0.0106 ± 0.0108 ± 0.0087 ± 0.0105 ± 0.0056 ± 0.0116 ± 0.0073 ± 0.0053 ± 0.0091 ± 0.0035 ± 0.0183 ± 0.0109 ± 0.0113 ± 0.0151
Vertebral column 0.5894 0.5914 0.6373 0.6061 0.6072 0.604 0.604 0.6468 0.6148 0.6348 0.6167 0.6249 0.6048 0.6095
± 0.0386 ± 0.0412 ± 0.0449 ± 0.0292 ± 0.0251 ± 0.0282 ± 0.0385 ± 0.0412 ± 0.0361 ± 0.0464 ± 0.0156 ± 0.0251 ± 0.0194 ± 0.0204
New thyroid 0.7741 0.7586 0.8003 0.8215 0.8317 0.792 0.2416 0.8057 0.7983 0.823 0.7842 0.7431 0.7684 0.7587
± 0.0335 ± 0.04565 ± 0.0190 ± 0.0339 ± 0.0222 ± 0.0421 ± 0.0385 ± 0.0499 ± 0.0311 ± 0.0205 ± 0.0521 ± 0.0690 ± 0.0707 ± 0.580
Dermatology 0.9089 0.9301 0.9559 0.9344 0.9434 0.9386 0.8412 0.9317 0.8889 0.8172 0.9115 0.9171 0.9024 0.9063
± 0.0150 ± 0.0091 ± 0.0068 ± 0.0170 ± 0.0100 ± 0.0233 ± 0.0321 ± 0.0119 ± 0.0204 ± 0.0176 ± 0.0141 ± 0.0135 ± 0.0206 ± 0.0173
Balance scale 0.2229 0.256 0.3643 0.2154 0.2083 0.2247 0.2516 0.2116 0.3379 0.429 − 0.2735 0.1405 0.1339 0.3307
± 0.0239 ± 0.0352 ± 0.0706 ± 0.0217 ± 0.0293 ± 0.0178 ± 0.0180 ± 0.0189 ± 0.0492 ± 0.0364 ± 0.0178 ± 0.0110 ± 0.0320 ± 0.0214
Glass 0.2494 0.301 0.3623 0.4125 0.2796 0.3909 0.2321 0.3011 0.0595 − 0.1694 0.2539 0.2713 0.1628 0.2529
± 0.0496 ± 0.0596 ± 0.0439 ± 0.0512 ± 0.0915 ± 0.0489 ± 0.0805 ± 0.0389 ± 0.0894 ± 0.0081 ± 0.0624 ± 0.0398 ± 0.0666 ± 0.0314
Heart − 0.4132 − 0.3978 − 0.3753 − 0.444 − 0.4106 − 0.4353 − 0.4389 − 0.3518 − 0.3794 − 0.3506 − 0.4077 − 0.3952 − 0.3878 − 0.3841
± 0.0261 ± 0.0269 ± 0.0278 ± 0.0327 ± 0.0376 ± 0.0393 ± 0.02901 ± 0.0429 ± 0.0357 ± 0.08237 ± 0.0301 ± 0.0476 ± 0.0302 ± 0.0324
Car evaluation 0.9731 0.8248 0.9697 0.9921 0.9701 0.6457 0.9843 0.8061 0.5592 0.9767 0.3903 0.6436 0.7439 0.8122
± 0.0166 ± 0.0172 ± 0.0085 ± 0.0070 ± 0.0139 ± 0.0230 ± 0.0072 ± 0.0208 ± 0.0273 ± 0.0057 ± 0.0382 ± 0.0084 ± 0.0122 ± 0.0160
Thyroid 0.9494 0.943 0.9569 0.9471 0.9431 0.9537 0.9358 0.9603 0.8774 0.7566 0.3333 0.9604 0.9561 0.3335
± 0.0044 ± 0.0049 ± 0.0065 ± 0.0062 ± 0.0065 ± 0.0023 ± 0.01143 ± 0.0038 ± 0.0306 ± 0.0351 ± 0.0000 ± 0.0049 ± 0.0029 ± 0.0004
Yeast − 0.0095 − 0.1297 − 0.0283 − 0.0228 − 0.0796 0.0415 − 0.0822 − 0.0437 − 0.2991 − 0.1124 − 0.2593 − 0.0542 − 0.0061 − 0.0483
± 0.0361 ± 0.0132 ± 0.0192 ± 0.0244 ± 0.0266 ± 0.0195 ± 0.0148 ± 0.0194 ± 0.0269 ± 0.0098 ± 0.0293 ± 0.0224 ± 0.0333 ± 0.0361
Page blocks 0.5489 0.4927 0.6272 0.5448 − 0.3273 0.572 0.5375 0.5423 0.4592 0.1063 0.4042 0.4137 0.5912 0.3993
± 0.0244 ± 0.0325 ± 0.0170 ± 0.0239 ± 0.0693 ± 0.0341 ± 0.0273 ± 0.0325 ± 0.0337 ± 0.0125 ± 0.0112 ± 0.0192 ± 0.0129 ± 0.0235
Shuttle 0.8751 0.9385 0.9176 0.9266 − 0.2663 0.9156 0.6635 0.9682 0.7336 − 0.1741 0.8661 0.873 0.8784 0.8767
± 0.0302 ± 0.0190 ± 0.0170 ± 0.0244 ± 0.0498 ± 0.0262 ± 0.0248 ± 0.0139 ± 0.0256 ± 0.0282 ± 0.0226 ± 0.0250 ± 0.0154 ± 0.0283
Average 0.5304 0.5226 0.5768 0.5564 0.4124 0.5411 0.4855 0.5420 0.4618 0.3766 0.3992 0.4739 0.5071 0.4831
  1. The best performance is shown in italic for each dataset