Skip to main content
. 2022 Aug 8;19(15):9756. doi: 10.3390/ijerph19159756

Table 2.

Comparison of models with and without the balancing technique.

Data
Balancing
Methods Metrics
Accuracy Sensitivity Specificity AUC
Mean (SD) Mean (SD) Mean (SD) Mean (SD)
Without balancing Lasso 85.70 (0.48) 0.00 100 (0.01) 63.00 (1.11)
XGB 84.73 (3.17) 3.08 (8.91) 98.35 (5.12) 63.22 (1.04)
With undersampling balancing Lasso 58.97 (0.94) 60.97 (2.03) 58.64 (1.19) 62.80 (1.05)
XGB 54.30 (16.56) 63.00 (15.38) 52.87 (21.83) 62.32 (1.25)
With oversampling balancing Lasso 58.23 (0.83) 60.67 (1.89) 58.89 (0.99) 62.66 (1.10)
XGB 25.62 (6.69) 87.65 (6.72) 15.26 (8.86) 59.26 (1.19)