Design And Implementation Of Phishing Email Detection System Based On Psychological Feature Analysis

Social engineering attack is a common means of network attack,and its main attack carrier is email,that is,phishing email attack.The ubiquity of e-mail use makes phishing attacks extremely harmful,and due to the continuous improvement of its camouflage methods,it has seriously threatened the information and property security of network users.At present,most phishing email detection algorithms use the number of occurrences of related symbols and whether keywords appear as the basis for feature extraction.The features extracted by this method are not accurate enough,which will directly affect the detection effect of phishing emails.To solve this problem,this paper proposes a phishing email detection algorithm based on psychological feature analysis.The features used in this algorithm include psychological features and original features.Psychological features include persuasive principle features and emotional features,and original features include email header features and email body features.The persuasive principle features are extracted by the persuasive principle feature extraction model based on BERT.Since there are few public data sets for the persuasive principle feature,this paper also constructs a persuasive principle feature data before training the persuasive principle feature extraction model.The persuasive principle features include the three-dimensional features of scarcity,reciprocation and authority.The emotional features are extracted by the emotional feature extraction model based on BERT and TextCNN,including pessimism,disgust,surprise,optimism,trust,joy,fear,sadness,expectation and the total number of emotional features.This paper uses the Stacking integrated training method to train a phishing email detection algorithm based on psychological feature analysis,and studies the influence of the selection of the base learner and the training method of the meta-learner on the final detection effect of the algorithm during the integrated model training.Finally,the effectiveness of the algorithm proposed in this paper is verified by experiments.This paper uses the Stacking integrated training method to train a phishing email detection algorithm based on psychological feature analysis,and studies the influence of the selection of the base learner and the training method of the meta learner on the final detection effect of the algorithm during the integrated model training.Finally,the effectiveness of the algorithm proposed in this paper is verified by experiments.Based on the phishing email detection algorithm proposed in this paper,combined with the blacklist detection method,a phishing email detection system based on psychological feature analysis is designed and implemented.This system mainly includes five parts:user management module,blacklist management module,email feature extraction module,phishing email detection module and detection result statistics module.In this paper,each module is functionally tested.The final system test results show that the system can effectively resist phishing email attacks.