Research On Anti-Fraud Web Page Based On User Feedback

In the 21 st century,the Internet is moving forward at high speed,and now has been affecting life every day,no matter encountered any problems to find the answer in the search engine has long been known in life.So that the requirements of the search engine is also getting higher and higher,everyone wants to enter the keyword in the search engine input box,the results of the page no longer need to dial the scroll bar can be found and match their own heart content.To do this,in addition to constantly digging the user's behavior information,to develop a more efficient similarity matching algorithm,on the other hand,to do with the fraud page for the game,to deep into the loopholes in the fraud page,to develop a new algorithm from the basis of fraud pages to curb the development of fraudulent web pages.The traditional anti-fraud web page algorithm has always been concerned about the characteristics of the web page itself,such as web content features,web link features and other related features,but ignored the largest group of the Internet,that is the user.Search engine development is for users to use,to study the new similarity matching algorithm and anti-fraud page is also for users to use the search engine more efficient.Since the traditional way to focus on the characteristics of the web page itself has no good results,it is better to convert to another way,allowing users to decide,through a mechanism,the automatic feedback generated information after users can visit the page,so you can get the user feedback feature data,use this data to dig out which page is a real fraud page,screening out a collection of non-fraud pages.When a user searches for a web page,whether it is a fraudulent web page or a non-fraudulent page,there are only two answers in the user's mind: match and no match.users can not only feedback both to achieve the effect of anti-fraud pages,but also for the personalized search algorithm to provide a little bit of support.This paper is anti-fraud web pages based on the user feedback,the purpose is through an algorithm can be in the user feedback feature data to dig out the page is a fraud page or not,thus screening out a collection of non-fraud pages.The work of this paper is divided into two parts:1 Three algorithms of dealing with user feedback feature data can be improved by reading the related literature such as fuzzy mathematics,clear theory,economics and statistics.Algorithm based on the fuzzy operator method,algorithm based on the novel matching clustering method,and algorithm based on the inverse of the positive index method,and they used in the public data set to verify the effectiveness and feasibility of the algorithm.2 By comparing the experimental results of these three algorithms,the algorithm based on the inverse of the positive index method can be used to deal with the user feedback feature data more reliably and efficiently,and compared with algorithm based on the characteristic attribute data of the web page itself.The results show that,The algorithm of anti-fraud based on user feedback is better,Accuracy,recall rate and F value all reached 0.99.