Research And Application Of Online Clickstream Fingerprints Recognition Algorithms

With the advent of the era of big data and the maturity of Internet technologies,especially the popularization,application and development of big data,Internet of Things and cloud computing technologies in various fields,the acquisition of human information is not limited to the physical space.The Internet is constructing a new social form.People browsing information,shopping,entertainment and office activities on the Internet every day have become an integral part of people's lives.The Internet space has become a virtual space parallel to the real space.The study of human online behavior in Internet space has rich scientific significance and application value.Based on the online behavior data provided by the China Internet Network Information Center,we studies the human online clickstream fingerprint recognition algorithm from the perspective of data science and classifies the user's heterogeneous population categories based on the proposed online clickstream fingerprint algorithm.A method for abnormal user detection based on online clickstream fingerprint is proposed.The main research content includes the following three aspects.Firstly,Research on online clickstream fingerprint recognition algorithm.Based on the user's online behavior data,a quantitative study of human click behavior in cyberspace is carried out.Using machine learning algorithms and through a large number of experiments,it is found that there is indeed a click stream fingerprint that can identify the user's individual identity in the virtual space.In this paper,we model and analysis the online clickstream fingerprint,propose the identification algorithm of the individual online users constructed by online clickstream data.Through the constructed behavior characteristics labels using the user's online clickstream data online time information,click software information and click content information,and combined with machine learning algorithms,we identify individual users,which shows that the average accuracy of recognition can reach more than 90%.Secondly,Research on heterogeneous population identification based on online clickstream fingerprinting.Based on the online clickstream fingerprint recognition algorithm proposed in this paper,the online users are studied for the identification of heterogeneous population categories and the individual behavior characteristics of heterogeneous populations in cyberspace are quantitatively analyzed.Studies haveshown that independent individuals in the same group in cyberspace have similar behavioral characteristics.Based on the online clickstream fingerprint recognition algorithm,it can identify its heterogeneous population categories.A large number of experiments have shown that the recognition accuracy rate can reach 89.83% under the six dimensions of user demographic information: social class,age,region,education background,gender and household registration.Thirdly,Abnormal user detection based on online clickstream fingerprints.We propose a method to detect online abnormal users based on the online clickstream fingerprint algorithm.Studies have shown that this method can achieve a good anomaly user detection effect based on less clickstream historical data and clickstream live data,with an accuracy rate of more than 99%.As the number of users required to detect continues to increase,its accuracy remains stable.