Research And Implementation Of Web Abnormal Detection System

With the increasing popularity of the Internet and the rapid development of web technology,while people enjoy the convenience brought by webpage information,abnormal webpage attacks are eroding people's privacy and property safe.In recent years,abnormal webpages have become increasingly destructive,more concealed,and more diverse,and it has become more and more difficult to be identified.It gets hard for ordinary users to judge whether the webpage is abnormal.The existing web abnormal detection methods often cannot detect large amount known types and unknown types of abnormal webpages.Feature extraction is a key step in detecting abnormal webpages.The characteristics of different types of abnormal web pages have been thoroughly studied.According to the purpose and means of attack of abnormal webpages,it is divided into offensive malicious webpages,deceptive webpages,and spam webpages.A novel multi-type abnormal webpage detection method is proposed,and a web abnormal detection system is designed based on this method.This main works of this thesis are as follows: 1.Three traditional web abnormal detection methods are researched,the merits and demerits of these web anomaly detection methods are summarized.On this basis,a multi-type webpage feature extraction method is proposed.This method divides abnormal webpages into three types according to the attack purpose,and extracts the features of the abnormal webpages according to the attack purpose and methods of the three types.The method of numerical padding and normalization is used to solve the problems of missing values and huge range of values in the process of feature optimization.The improved SVM-RFE algorithm is used to eliminate redundant features.A SVM with feature weighted degree is designed to detect abnormal webpages.At the same time,the method is simulated on four public data sets and compared with other existing methods.The results show that the multi-type webpage abnormal detection method proposed in this thesis has higher accuracy.2.A web abnormal detection system is designed based on the multi-type webpage abnormal detection method.There are four modules in this system: the first is a web abnormal detection training module.A multi-type abnormal webpage detection method proposed in this thesis is used to implement a persistent web abnormal detection model.The second is abnormal webpages detection module.A multi-type abnormal webpage detection algorithm is implemented based on the Java language.The input is a feature vector made by vectorizing the URL entered by the user.The output is a detection result of 1 or 0,indicating whether it is an abnormal webpage,and the URL and detection results are stored into database.The third is the visualization module.Use the Bootstrap plugin to show information such as detection result information,detection trend information,and user-built blacklists in charts and other figure.The fourth is the system management module.This module contains user registration and login operation,user and system information management.The registration and login method uses email and login verification codes to ensure the safety of users when they use the login and registration functions,management of user and system information ensures that the system can operate accurately,securely,and efficiently.The functions and performance of the system are tested and the results show that the system has better robustness and load capacity,and better detection capability for abnormal webpages.