User Mobility Analysis And Application Based On Mobile Big Data

In the past few years,the number of mobile phone users in China has increased significantly.According to the"Statistical Bulletin of the Communications Industry of 2018" issued by the Ministry of Industry and Information Technology,by the end of 2018,the total number of mobile phone users of the three major operators in China has reached 1.57 billion.At the same time,the types of mobile communication services are constantly moving towards diversified goals.With the advent of the 5G era,traditional mobile communication networks cannot meet the needs of next-generation networks,and there is room for improvement in terms of delay,regional service capacity,transmission rate,and resource utilization.Limited by the system resources and the existing network architecture,we need to make full use of the massive mobile phone data generated by cell phone users,helping us conduct intra-regional mobility research and behavioral pattern analysis,and thus we can use mobile analysis and prediction results to solve the bottleneck problem of system resource utilization in 5G networks.The above problem can be effectively solved by using a resource scheduling scheme based on user mobility prediction.Using the big data to model the user's mobility,analyse the spatial and temporal distribution of mobility,and then summarize the user's frequent mobility patterns through improved mining algorithm.By establishing the RNN model based on spatio-temporal characteristics,the user mobility is predicted.And then apply it to the 5G network slice resource scheduling scheme.The main contents of the study are as follows:Firstly,in order to explore the regular characteristics of user's movement,providing theoretical support and reference for mobility prediction,the commonly used user mobility analysis method is applied to analyse user's spatial distribution,moving distance,radius of rotation,and dwell time from the perspective of group and individual.The results show that the moving distance and the radius of the radius of the group users are approximately consistent with the power law distribution with exponential truncation,and the individual's moving characteristics are not consistent.Secondly,in order to further explore the movement law of individual users,the coverage areas of different base stations are functionally divided combined with the information of POI,so that the dense urban areas are divided into different functional areas,so as to semantic modeling the user's movement trajectory.Then,an improved mining algorithm is used to mine the frequent movement patterns of individual users in different functional areas,which further verifies that the group users' mobility has certain commonalities at the macro level.Finally,in order to solve the bottleneck problem of system resource utilization in 5G network,the spatio-temporal featured RNN is used to predict the user's future location and to design 5G network slice resource scheduling scheme.Experiments show that compared with other schemes,this scheme can effectively improve the utilization of system resources.The work in this thesis has certain value for communication network deployment,municipal planning,and traffic flow control.