Table 8 Pairwise comparison of mean F1-scores across various different models on PIMA Dataset.

From: Optimizing imbalanced learning with genetic algorithm

Model

SMOTE9

ADASYN11

GAN71

VAE63

SGA

EGA

SVMGA

SMOTE

65.90

− 0.66 (0.4003)

− 0.98 (0.5447)

− 1.80 (0.2184)

− 1.62 (0.4308)

− 1.18 (0.1389)

− 3.88 (0.0451)

ADASYN

0.66 (0.4003)

66.56

− 0.32 (0.7748)

− 1.14 (0.3226)

− 0.96 (0.5750)

− 0.52 (0.1498)

− 3.22 (0.0146)

GAN

0.98 (0.5447)

0.32 (0.7748)

66.88

− 0.82 (0.6758)

− 0.64 (0.8073)

− 0.20 (0.8758)

− 2.90 (0.0722)

VAE

1.80 (0.2184)

1.14 (0.3226)

0.82 (0.6758)

67.70

0.18 (0.8626)

0.62 (0.6077)

− 2.08 (0.1078)

SGA

1.62 (0.4308)

0.96 (0.5750)

0.64 (0.8073)

− 0.18 (0.8626)

67.52

0.44 (0.8007)

− 2.26 (0.1784)

EGA

1.18 (0.1389)

0.52 (0.1498)

0.20 (0.8758)

− 0.62 (0.6077)

− 0.44 (0.8007)

67.08

− 2.70 (0.0615)

SVMGA

3.88 (0.0451)

3.22 (0.0146)

2.90 (0.0722)

2.08 (0.1078)

2.26 (0.1784)

2.70 (0.0615)

69.78

  1. Diagonal entries show the average F1-score for each model. Off-diagonal entries represent the mean difference in F1-score between the row and column models, followed by the p-value from a paired t-test in parentheses.
  2. Significant values are in bold.
Back to article page