User Behavior Analysis Based On Mobile Internet Data

As the mobile Internet developed rapidly,the data has already become an important link between society and the Internet.The analysis of users' behavior based on mobile Internet data is helpful to reveal the dynamics of users' behavior and analyze various social phenomena.The research results can greatly enrich the research and application of social behavior.Under the background of the mobile Internet,this paper makes full use of mobile Internet detailed record(UDR)data to study various user behaviors,and then infers user behavior characteristics to provide the technical support to telecommunication business operations,network management,and application recommendation,which also provides a new approach for the research of social behavior.This paper selects three representative aspects to conduct research,including the user APP access behavior,user social APP access behavior,and traffic model generated when users surfaces the Internet.The widespread APP utilization is the most significant feature of the current mobile Internet,and the study related to the user APP access behavior can discover user content preferences,user APP access characteristics,and people's behavior patterns in the virtual space of the mobile Internet,which can be employed in business development and social management.Social APP is the most commonly APP in daily work and life,and it is also the most commonly used one.However,it also results in the biggest impact on traditional telecommunications services at the same time.This paper first quantitatively studies the impact model of social APP on traditional call services,and classifies and mines the impact of social APP access behaviors on call services of different user groups,and this work provides a reference for the research on the mapping model of online behavior and offline behavior.With the rapid growth of mobile Internet traffic,how to analyze user traffic characteristics and to predict traffic is one of core issues that attracts great attention from the telecommunications operators,and reflects the user traffic usage patterns.This paper studies the user space traffic model from the perspective of base stations.The main research work of this paper is organized as follows:1.User access behaviors of various APPs.This article uses a data-driven method to analyze the user's various APP access behaviors.On one hand,it reveals the preference characteristics and basic composition patterns of user access content,and discovers the behavior pattern of the mobile Internet virtual space.On the other hand,it can help mobile Internet companies and mobile phone manufacturers to improve operating efficiency.This paper proposes a context-based user APP access pattern discovery method,which can solve the problem of large-scale user content preference feature extraction.By utilizing the user's records of multiple APP visits and a polymorphic model,this method builds the context dynamic process of user APP visits as a user APP access transition probability tensor.The tensor is extracted for the high-dimensional features,which reveals the typical characteristics of APP access behavior of user groups.The user APP access behavior is expressed as a combination of typical features,and is approximated by the typical characteristics and weights.This method combines data from Beijing,Jinhua,Zhejiang and Jiangmen,Guangdong,and analyzes in multiple dimensions.It reveals the characteristics of APP access behavior of user groups in terms of different cities,different brands,and different times.2.User social APP access behavior.Social APP is currently the most common one visited by users.Unlike the information-recommended APPs,such as news and e-commerce,social APP have extremely strong interaction functions.Therefore,in the field of social behavior,there still exists a big debate about "Is there an impact on the business? How difference is the impact on different groups? ",and it has become a problem that telecommunications operators need to explore urgently.This article first quantitatively analyzes the impact model of the social APP conversation voice call service,proposes a user social software and voice call causality reasoning method based on the tendency score matching method,and subdivides the user group according to age and gender to further study the impact of different groups and different social software on user voice calls.These research results show that the influence of social APP conversation voice calls is not just a simple negative effect.Utilizing the social APP of the specific user groups will promote the use of voice call services and increase the volume of voice calls.Considering both popular social APPs including the We Chat and the QQ,this paper adopts the tendency score matching method to analyze the impact on different user groups.This study provides important reference for the study of users' online and offline behaviors and their correlations,and has a positive significance for telecommunications operators to optimize the design of business packages and tap the potential of users.3.The user space traffic model and its application in traffic offload strategies.Although the network communication equipment is continuously upgraded,the growth of network capacity still lags behind the increment of the users' demands for network traffic,and this lagging trend is becoming more and more serious.Therefore,it is necessary to analyze the distribution of traffic in the time and space domains to find the characteristics of user traffic and formulate corresponding traffic offloading strategies.This paper proposes a base station traffic prediction model based on spatial cooperation,which converts the user traffic into base station traffic,and uses the base station spatial traffic cooperation relationship to improve the accuracy of target base station traffic prediction and to offload the traffic.Thus,the users' requirements can be guaranteed,and the mobile traffic load pressure can be reduced by making full use of WIFI,with maximum of system throughput.Based on the UDR data,this paper studies and analyzes user behavior from multiple aspects.On one hand,it proposes a variety of user behavior analysis methods and finds user behavior characteristics in multiple dimensions,which is helpful to the study of social behavior.On the other hand,it taps the potential value of UDR data which provides the support for mobile communication network management and Internet applications.