Design And Implementation Of Ananti-Phishing Detection System Based On Expert Knowledge Database

Phishing is a phishing attackers frequently used common network of people’s daily life.as carrier. They often elaborate design the target sites are very similar to phishing sites as bait through the network carrier by spam, instant messaging, social networks, mobile phone short message and so on, however, user didn’t realize their information has been stolen, such as user’s account information, password, or credit card details has been stolen.Phishing become more and more rampant in recent years around the world, it, is serious influence to people’s normal work and daily life. Therefore, this paper based on the analysis of the phishing technology of the Predecessor’s works. This paper proposed an anti-phishing system based on expert knowledge database. It takes advantage of the merits of knowledge database. What’s more, it improves the knowledge database to design the anti-phishing system based on expert knowledge database. From the phishing URL structure and features of the page as the foundation, the system is made of blacklist\whitelist knowledge database、features of URL knowledge database and features of page knowledge database based on previous studies. Through mass data and Detailed experimental, the system got a good result.In the following, we summarized this paper’s contributions to the problem of detecting phishing attacks:1、It proposed an anti-phishing system based on knowledge database. It used the characteristics of knowledge database, such as simply、fast and so, meanwhile, it designed the characteristics of URL knowledge database and characteristics of page database to solve some drawbacks of the knowledge database.2、The system analyzes the characteristics from the phishing URL The features of URL and matching rules have been stored in the knowledge database. If there is a new feature of URL or algorithm this system could update to the knowledge database. This paper used logistic regression as features machine learning classifier to detect the phishing sites that relies on these selected.3、The system analyzes the characteristics from the phishing page. It combined DOM features of phishing page, and then presents eight kinds of phishing pages feature. It uses support vector machine algorithm implementing the feature pages of classified.4、It proposed a website ALEAX ranking, ICP filing and domain name registration information identifying the site to make supplementary judgment in the final part besides database.Previous anti-phishing technologies only took consideration of features of URL or features of page, the updated technology is also lack. By contrast to these technologies, this paper take consideration of features of URL and features of page, and it take advantage of knowledge database to store features and matching rules, after all it achieves a better performance.