Research Of Detection Methods On Click Fraud In Network Advertising

With the rapid development of Internet technology and Internet users increased,the network has become a more suitable for the dissemination of advertising and get better display vector. As the rapid development of online advertising has brought unlimited business opportunities for advertisers and Internet sites at the same time, the network has become one of the main economic sources of the major Internet Co.Current domestic and international researches on the click fraud detection are mainly divided into three directions: the improvement of payment mode, the verification code prevention and data analysis. However, for click fraud detection, advertisers are still difficult to accept changes in the current pay per click business model. Existing methods are difficult improve these problems.On the premise of not changing the charging mode, this paper draws on the existing problems and puts forward a more reasonable solution to the existing problems. The click fraud detection model mainly includes three parts: data acquisition, user identification and analysis. The data acquisition module collects the browser fingerprint and EverCookie storage data through the client code. The collected data is used for user identification. At the same time, the data acquisition model will be collected to click on the relevant data to detect fraud. Identification module mainly through the browser fingerprint identification technology and EverCookie technology to identify the user's identity. Parallel fraud detection module is divided into parallel statistical analysis and parallel suspicious analysis, based on the Hadoop distributed environment for parallel analysis and detection of user click dataAt the end of this paper, the click fraud detection model is verified. The results show that the proposed algorithm has good detection effect in the click fraud. In terms of efficiency and accuracy of click fraud detection has been improved obviously, and be able to conduct a reliable detection of click fraud in online advertising.