User Precise Positioning And Behavior Analysis Based On Big Data

With the rapid development of the mobile Internet,human beings have entered the era of "big data".The arrival of the era of big data has brought both tremendous value and severe challenges.The increase in user demand for location services has led to more and more attention to user location technology.Although there are GPS positioning,base station positioning and so on,these methods have some problems,such as some methods are affected by the environment easily,and some methods need to consume a large amount of resources.At the same time as the rapid rise of mobile networks,various services for users are also emerging.How to stand out in the fierce market competition and provide users with better service has become a top priority for various service providers.One of the hotspots is to analyze the behavior of mobile users,tap the habits of users,understand the preferences of users,and provide users with personalized services.Therefore,in today's huge data volume and rapid increase in service demand,it is very urgent and necessary to study the precise positioning and behavior analysis of users based on big data.The main work of the thesis includes:Firstly,in the accurate positioning of users based on big data,the automatic construction of location fingerprint database is studied,including automatic extracting features of MR(Measurement Report)data;automatic mining of GPS location information in data services;automatic associating MR data feature with GPS location information.By studying the location matching algorithm to mine the similar population,the approximate location of the user to be located is obtained.And,the road matching algorithm is used to match the user location to the road closest to its approximate location,so as to accurately locate the user's location.The algorithm was verified by using 14784 effective position data,and the results show that the proportion of positioning error below 100 meters is 61.7%.Secondly,in the aspect of big-data-based user behavior analysis algorithm,the long short-term memory network LSTM(Long Short-Term Memory)is used to construct a model describing user behavior to achieve more accurate reflection of user traffic consumption.At the same time,in order to overcome the over-fitting problem,Dropout was introduced.The algorithm is verified by using real data,and the results show that 90%of the 500 users who are scaled out have a mean square error of 13,805,465 bytes(13MB)or less per hour.Finally,the visualization system was presented.The visualization system includes trajectory visualization platform and traffic visualization platform.Among them,in the traj ectory visualization platform,on the basis of obtaining the precise position of the user,the researcher can use the visualization platform to import the user position,and the visualization platform automatically draws trajectories of the user according to the chronological order of the user position point,and then researcher can use the trajectory visualization platform to observe and analyze the user's movement trajectory.In the user traffic visualization platform,researchers can intuitively observe the user's traffic consumption,analyze the user's behavior patterns,business preferences and so on.The research results of this paper can help operators reduce the resources needed for user positioning and provide users with a better experience,at the same time,it can also help optimize the distribution of business outlets and the setting of base stations.On the other hand,it can help service providers understand user needs better,improve their competitiveness,and improve user satisfaction.