Mobile Internet User Clustering Based On The Analysis Of URL

In recent years,the development of mobile internet is rapid.As the data CNNIC publishedon July2012shows,60%of people access the Internet through mobile devices.Alarge numberof mobile Internet users is indicating that the mobile Internet market has great potential.Chinese telecom operators attache great importance to mobile Internet market,they areactively analyzing mobile Internet user’s online behavior.Base on user behavior,the operatorscan cluster different customers,identify target customers,provide users with differentiatedservices,or send messages the users may be interested.Under this background,this paper studies and implements a Web page classificationalgorithm based on URL.This algorithm uses the characteristic frequency for web URLclassification.Because of its low computational complexity,it is a fast web classificationalgorithm.Using this algorithm, the paper implements a clustering system of mobile Internet usersbased on URL analysis. The system cluster users according to their online trajectory,and findthe target customers by setting the marketing rules,then complete the precision marketing totarget customers.The research work mainly includes the following aspects:First, the paper studies the traditional webpage classification algorithm and classificationalgorithm.The paper proposes a classification algorithm based on the URL webpage.Thealgorithm can classify webpage only by using the characteristics of URL webpage.The mainadvantage of this algorithm is its speed.Second，according to a CMCC province branch marketing needs,the paper designs amobile Internet user clustering system.The system analyzes DPI (Deep Packet Inspection) log,using the proposed classification algorithm based on the URL of pages,and then clustering formobile Internet users.The system is directly connected to the advertising distribution system,which can achieve precision marketing to the target customers.Third， the paper proves the functionality of the system through a marketingexamples,and analyze the the system’s efficiency on precision marketing.