A Malicious URL Detection Model Based on Convolutional Neural Network

Table 1

Experiment results.

No.	Embedding method	Network structure	Accuracy	F1-score	Recall	Precision
1	Character embedding	Stacked CNN	0.923	0.926	0.964	0.890
2	Word embedding	Stacked CNN	0.954	0.953	0.931	0.977
3	Word embedding based on character embedding	Stacked CNN	0.958	0.959	0.976	0.942