Skip to main content

Table 3 Average F1 scores on different K values. The values in brackets represent the standard deviation (Std) in 5-fold cross validation. The proposed SASMOTE algorithms perform better on the average F1 score with respect to the value of K ranging from 10% to 75% while SMOTE performs better with respect to the value of K = 5%

From: A self-inspected adaptive SMOTE algorithm (SASMOTE) for highly imbalanced data classification in healthcare

Model

\(\boldsymbol{K = 5\%}\) (Std)

\(\boldsymbol{K = 10\%}\) (Std)

\(\boldsymbol{K = 25\%}\) (Std)

\(\boldsymbol{K = 50\%}\) (Std)

\(\boldsymbol{K = 75\%}\) (Std)

SASMOTE

49.15 (9.02)

47.57 (9.80)

51.38 (8.49)

51.30 (5.99)

51.16 (6.32)

SASMOTE w/o visible

49.71 (9.79)

49.69 (9.82)

49.95 (7.58)

48.11 (7.76)

49.72 (9.67)

SASMOTE w/o inspection

46.81 (9.63)

48.57 (6.97)

50.38 (8.64)

51.38 (12.29)

50.26 (11.94)

B-SMOTE

48.84 (9.20)

48.63 (7.61)

48.02 (8.50)

46.14 (7.43)

47.80 (8.12)

SMOTE

49.73 (10.09)

47.84 (9.30)

50.15 (10.49)

48.69 (6.58)

46.07 (11.19)