With the development of Internet technology, network security is facing great challenges. Malicious URL detection can defend against attacks such as phishing, spams, and malware implantation. However there are some problems on current malicious URL detection, for example the methods used to extract features are inefficient and hard to adapt to the current complex network environment. To solve these problems, this paper uses the word embedding method based on character embedding as the way of vector embedding to improve the deep convolutional neural network, and designs a malicious URL detection system. Finally, we carry out experiments with the system, the results prove the effectiveness of our system.
Paper
Full text
A Malicious URL Detection Model Based on Convolutional Neural Network
Semantic Scholar · Computer Science · 2020
Abstract
With the development of Internet technology, network security is facing great challenges. Malicious URL detection can defend against attacks such as phishing, spams, and malware implantation. However there are some problems on current malicious URL detection, for example the methods used to extract features are inefficient and hard to adapt to the current complex network environment. To solve these problems, this paper uses the word embedding method based on character embedding as the way of vector embedding to improve the deep convolutional neural network, and designs a malicious URL detection system. Finally, we carry out experiments with the system, the results prove the effectiveness of our system.