Research On Fingerprint Identification For Specific Website

Nowadays, with the deepening of the privacy protection technology, a variety of anonymous communication technology including VPN, SOCKS proxy and Tor has been developed quickly in recent years.Because the communication content and communication sides are hidden, it is difficult to rely on the past detection means based on the contents of the packet to monitor the Internet.Therefore the researchers proposed a website fingerprint recognition technology based on passive traffic analysis. In fact, we often need to determine whether the target has visited a particular website, however the website fingerprint identification system cannot solve this kind of problem properly; at the same time, in the past research, website fingerprint data collection and fingerprint training set may not be considered comprehensively, which greatly reduce the applicability of fingerprint recognition system.On the basis of previous research findings, aiming at the above problems, this thesis puts forward a new web page fingerprint identification system. Compared with the previous web page fingerprint identification system, this system has strengthened the fingerprint identification on a particular website’s page, which greatly improves the recognition rate of particular website’s page fingerprint. This thesis studies how the web browser cache technology impacts on the web page fingerprint recognition in deep, fully considered the difference of fingerprint data when using different browsers, which ensures the reliability of the fingerprint training set under complicated conditions; We handle the classification imbalance when recognize specific website’s page by classifying the training set and classification result integration. At the same time, in order to reduce the computational complexity of the similarity classifier, for the first time we classify the data using feature selection algorithm,which furtherly improved the performance of fingerprint classification;Finally, this thesis compares the performance difference between the previous web page fingerprint identification system and new fingerprint identification system in this thesis, and furtherly clarified the superiority of the website’s page fingerprint identification system.