Table 1.

Average evaluation metrics of the two deep learning models and the five traditional machine learning classifiers. The relatively best two metrics in each column are shown in bold.

Classifier	Accuracy	Recall	Precision	F1-score	AUC
ResNet	0.8844	0.9325	0.8623	0.8952	0.9248
ResNeXt	0.8784	0.8944	0.8867	0.8905	0.9070
SMO	0.8082	0.9216	0.7607	0.8286	0.8623
Linear Regression	0.7606	0.8328	0.7403	0.7796	0.8258
Random forest	0.7314	0.8031	0.7148	0.7529	0.7895
Bagging	0.7113	0.7751	0.6999	0.7339	0.7514
Multi-layer Perceptron	0.7827	0.8552	0.7523	0.7971	0.8331