Click Fraud Detection In On-Line Advertisement And Realization

The click fraud is not only complex but also difficult to detection. The detection technology is a new research in the internet area. Expert system is the most remarkable branch these years during the artificial intelligence coming to the practicality research. Using expert system to achieve click fraud detection has accurate and efficient features, and for continuously improve the detection knowledge, the expert system has the flexibility to improve detection methods.The paper first summarized click fraud in CPC (Cost Per Click) business model at on-line advertisement. After analysis traditional click fraud detection schemes which are based on IP and Referer monitor, we provide an enhanced detection scheme in real time which is based on deterministic finite automaton (DFA).Then, we define its layered service architecture, introduction of the expert system and go into the details of the working mechanisms of rule engine. Also, we are realizing the detection rule with Drools which is an open-source project of rule engine. Finally, we use a case to study the comparison of the traditional detection methods and enhanced detection methods.