Detecting Phishing Emails Based On Authentication And Classification Integration

In recent years,the rapid development of the Internet has become an indispensable presence in people's daily lives.Various network security risks followed.There has been a cybercrime type APT(Advanced Persistent Penetration Attack)for commercial and political purposes.Common penetration breakthrough methods include e-mail,instant messaging,and website hanging horses.Phishing email is an infiltration attack through email.It is a very common attack method.Now there is a new form of phishing email,spear-phishing email,which achieveds phishing by disguising as acquaintances of the targets.However,each individual's personality,gender,and writing habits are different.Even if they are deliberately imitated,they cannot be completely similar.Therefore,this paper uses a sender authentication method to detect spear phishing emails.In verifying the sender's identity,style,gender,and personality traits need to be extracted from the email,and then the classifier is used to classify the email.However,ordinary phishing emails do not have the feature of disguising senders as people who are not familiar with the target.Therefore,verifying the sender's identity is not suitable to detect all phishing emails.For other cases,phishing emails can only be detected using common detection methods.This paper separately uses URL features,keyword features and email body features to classify and construct base classifiers,and then integrates the results by voting to obtain the final classification results.Finally,more features are extracted during the detection process,which can lead to overfitting.Therefore,in this paper,a dimension reduction method based on density and distance is proposed.The method can effectively reduce the dimension of the original feature space to a simple and representative two-dimensional vector by calculating the distance and local density between features.In this way,it can achieve the effect of saving time and improving accuracy and other performance.The final experimental results show that the phishing detection method based on sender authentication and classification integration studied in this paper is effective.