Table 2 The mean cross-validation scores, standard deviation (SD), and the accuracy from different algorithms for the SMOTE and original datasets. The table illustrates the mean K-fold cross-validation scores the corresponding SDs acquired by each classification algorithm with and without the application of the SMOTE.

From: Machine learning-based classification of histological subtypes of invasive breast cancer using MRI contralateral breast texture features

 

Algorithm

Mean accuracy

Standard deviation (SD)

Accuracy as percentage

With SMOTE

KNN

0.6121

0.0340

61.21%

LR

0.6440

0.0360

64.44%

RF

0.8723

0.0209

87.23%

DTC

0.8175

0.0197

81.75%

GaussianNB

0.6914

0.0315

69.14%

SVM

0.6833

0.0414

68.33%

AdaBoost Classifier

0.7356

0.0225

73.56%

Without SMOTE

KNN

0.7418

0.0185

74.18%

LR

0.7761

0.0034

77.61%

RF

0.8989

0.0224

89.89%

DTC

0.8653

0.0193

86.53%

GaussianNB

0.7508

0.0193

75.08%

SVM

0.7720

0.0029

77.20%

AdaBoost Classifier

0.8183

0.0210

81.80%