Design And Implementation Phishing Detecting System Based On Active Detection

Phishing is an internet term, which means criminals create false webpage as ameans of deceiving users into providing their credentials for financial benefit. In thepast year, nearly6million users visited different types of phishing websites each day.If users didn’t take care of webpages they visited and fill privacy information, theymight be exposed to economic injury. Detecting and blocking phishing websites intime has positive meaning for users property safety.As anti-Phishing technology developing, criminals use more methods for hiddencontent. In this paper, we use the page after browser rendering for detecting in orderto find hidden content. Because phishing detecting is a complex problem, we focuson web spoofing detecting that means the phishing webpages are similar to someofficial websites. For active detecting, we consider automatic generating URL fromprotected website, meta search, email and other date source.A visual block selected algorithm of webpage is proposed in this paper, whichcan split a full webpage into a set of independent visual blocks. Phishing and officialwebpages similarity calculation is based on visual blocks matching degree. Images inpages are also considered in visual block matching. A revised image perception hashcalculation method is used to pages quick compare which maps picture to a32bithexadecimal number.Support Vector Machine is used to classify content type of visual blocks, andNaive Bayes algorithm applies to judge a page either safe or phishing. A virtual blocksindex is built for speeding matching time. Except for protected website list, black listis used in detecting system. Black characteristics are intent to raise recalling rate, sowe can detect the phishing webpages built from same template.In conclusion, an initiative phishing detecting system is designed at last. Byinputting protected and phishing list, system builds white and black characteristicsdatabase. Experimental results show our method can find web spoofing and pagebuilt from black template.