Analysis And Modeling Of Mobile Internet User's Behavior In Multiservice Environment

With the rapid development of mobile Internet,intelligent terminals,such as smart phones and wearable devices,have broken through the network boundary and became an important entry and exit of Internet data.They are fully integrated into many fields such as education,automobile,medical treatment,finance,tourism,service etc.Meanwhile,the users in the field of mobile Internet applications have specific characteristics,behaviors and preferences.Using their characteristics to characterize the users accurately and achieve the refined operations has become the core competence of the operating enterprise in the mobile Internet business development.However,in the mobile Internet environment,there are many features of big data such as multi-source heterogeneous,widely dispersed,pattern coming after the data,etc.This brings more challenges to analyze the user behavior in the mobile Internet environment.Thus this thesis analyzes and models the mobile Internet users' behavior in the multi-service environment such as voice service,short message service and network flow service in telecom industry.The major work of the thesis is listed as follows:First of all,for the multi-service environment,massive unstructured data was generated when users use the mobile network resources.And the relevant data are preprocessed by protocol analysis and service characteristics to improve the quality of data for the next step in modeling analysis.Secondly,based on the research of the existing user portrait modeling method,the thesis proposes a method of user portrait modeling based on density cut K-means algorithm aiming at the problem of lack of refined mining analysis and unilateral user attribute analysis.On the basis of the user fact labels,this method extracts the user hidden labels by optimizing the K-means algorithm to describe the user behavior characteristics comprehensively.The empirical research shows that the modeling method can effectively extract the hidden information of users,fully reflect the potential demand of customers and provide the possibility of precision marketing.Thirdly,for the different attributes of users are rarely association analyzed combined with multibusiness environment,a business-based support calculation method is proposed.Based on the analysis and processing of different service data,the Apriori algorithm is used to analyze the moving attribute and behavior attribute of the user on the Hadoop platform.The empirical research proves that there is a correlation between the user moving attribute and different business environments.The research work shows that establishing the user portrait model and user behavior analysis model can provide quantitative theory support for the telecom enterprises and the basis for the market strategy choice to improve the efficiency and benefit of the market strategy.It is important in theoretical research and practical application.