Malicious URL Detection Method Based On Deep Learning

With the continuous development of the Internet,people's daily life has formed an inseparable relationship with the Internet.From the very beginning of socializing,browsing the web to now shopping,paying bills,etc.,all can be done on the Internet,but this also allows malicious attacks.Currently,Web attacks are constantly evolving,and many network applications have exposed various vulnerabilities and are facing various network attacks and threats.In traditional methods,detection methods based on blacklists and whitelists are used more,but their shortcomings are gradually exposed.This puts forward higher requirements for the accurate detection of network attacks.Because deep learning has its own characteristics,its use in the detection of malicious attacks has become one of the hot research fields.The URL detection method studied in this paper is mainly divided into two stages.First,by learning the existing malicious URL detection methods,understanding the current domestic and foreign research results in related aspects.On this basis,combined with Kolmogorov complexity theory,the data set is processed in the first stage.In this stage,a preliminary classification of the data set is carried out through data compression technology.Experiments have proved that the data processing at this stage can effectively improve the accuracy of later model detection.Secondly,this article introduces the method of deep learning to process the URL in the second stage.In this stage,combined with fast Text word embedding technology,the proposed deep learning model is used to train and predict the data,and at the same time to compare with the trained results of other models.The experimental results show that the deep learning model proposed this time has a good performance.In the last step,by combining the methods of the first stage and the second stage,on the basis of ensuring the accuracy,the training and detection efficiency of the model is further improved.Experiments prove that this method has good practical value.