Skip to main content
. 2022 May 19;23:189. doi: 10.1186/s12859-022-04715-w

Table 1.

Comparison of the evaluation metrics between MAGCNSE and its four variants

Method Accuracy Sensitivity Specificity Precision F1-score MCC
MAGCNSE-fgl 0.9029 0.9013 0.9043 0.8984 0.8998 0.8056
MAGCNSE-natt 0.9013 0.9068 0.8959 0.8952 0.901 0.8026
MAGCNSE-nattcnn 0.8885 0.9003 0.8783 0.8647 0.8822 0.7771
MAGCNSE-ncnn 0.9013 0.896 0.907 0.9128 0.9043 0.8025
MAGCNSE 0.9395 0.9192 0.9626 0.9654 0.9417 0.88

The bold number is the highest value of each column and its clarifies the superiority of our model