Skip to main content
. 2021 May 26;124:119–132. doi: 10.1016/j.future.2021.05.019

Table 6.

Comparison of the seven state-of-the-art models based on classification metrics on the test dataset of both Dataset 1 and Dataset 2.

Metrics Models
VGG16 InceptionV3 ResNet50 VGG19 MobileNet InceptionResNetV2 DenseNet121
Dataset 1

Accuracy 1.000 0.98 0.945 1.000 1.000 0.995 1.000
Precision 1.000 1.000 0.978 1.000 1.000 1.000 1.000
Recall 1.000 0.960 0.910 1.000 1.000 0.990 1.000
f1-score 1.000 0.979 0.943 1.000 1.000 0.994 1.000
Cohens Kappa 1.000 0.960 0.890 1.000 1.000 0.990 1.000
ROC accuracy 1.000 0.980 0.945 1.000 1.000 0.995 1.000

Dataset 2

Accuracy 0.987 0.994 0.952 0.991 0.992 0.994 0.990
Precision 0.980 1.000 0.954 0.990 1.000 1.000 0.996
Recall 0.996 0.990 0.952 0.994 0.986 0.990 0.986
f1-score 0.988 0.995 0.953 0.992 0.986 0.995 0.991
Cohens Kappa 0.988 0.989 0.905 0.983 0.985 0.989 0.981
ROC accuracy 0.987 0.995 0.952 0.991 0.985 0.995 0.991