Analysis And Research On Behavior Of Internet Users For Big Data

With the rapid development of information technology, information has penetrated into allaspects of live and work. The information explosion brings massive data. Data is expandingrapidly, and it determines the future development of the enterprises. With the passage of time,people will become more and more aware of the importance of data for the development of theenterprises. The great value of big data is gradually recognized by people. Through technicalinnovation and development, as well as the overall perception, data collection, analysis, sharing, itprovides a brand-new method for people to view the world. Therefore, how to use massive data,and to found the patterns, has become an important subject.Network operators provide services to the internet users, meanwhile they would store users’log data. Through these data, the operators can get the behavior characteristics of users. Theyadopt different marketing methods for different users, so as to promote the development ofenterprises.Therefore, this paper takes the access log data of a network operator as the foundation. By theanalysis of the mining, we can get characteristics of the users, and classify users according to theattributes into high and low consumption tendency. In this paper, the results of the work are asfollows:(1) Given the system design of analysis on internet users’ behavior for big data. This systemis composed of web log preprocessing, users’ feature extract and users’ behavior classification.This paper introduces the design process of each module in detail.(2) Given the method of users’ feature extract based on MapReduce. This paper takes user’sbrowsing goods category, residence time, frequency, geographic location, history and so on asusers’ behavior feature, meanwhile introduces the parallelization of this method.(3) Given the design of a naive bayesian classifier based on weighted attributes. According to the users’ feature, this paper classifies user’s behavior by using weighted attributs classifier, anddesign the classifier based on MapReduce.(4) In the environment of Hadoop this paper has implemented the internet users’ behavioranalysis system. According to the system, it can extract the attributes of the users, and based onthe classifier, it’s easy to classify the test users into high propensity to consume or lowconsumption tendency.