Design And Implementation Of Mobile Terminal Information Acquisition And Analysis System Based On Location

In recent years,with the rapid development of mobile Internet and the rapid spread of intelligent devices,it is very convenient to get the user's location information.These location data usually contain a lot of information,which for the analysis,prediction of human behavior model has a very important role.In this thesis,the main research direction is user's stay point and location prediction.Using Android system and Web development technology to develop projects.We successfully complete location-based mobile terminal information collection and analysis system.Firstly,the thesis reviews the background and significance of the research related to the subject,and introduces some achievements in the field of location data mining.Then some of the technologies involved in the project are briefly explained,including the Android system,Android's basic components and multithreading technology,several kinds of stop point recognition methods and commonly used position prediction technology.Followed by the development of the system design,and in accordance with the program client and server code preparation work.The client of system based on the Android system,the server based on Servlet,MySQL implementation,Web display using JavaScript language.The main function of the client is to collect the user's location information.The main function of the server is to store,display and analyze the user's location information.Secondly,the thesis introduce the module of stay point extraction.Stay point is a place where a user takes a while to move.The purpose of the stay point extraction module is to identify the stay points from the user's history track.In view of the fact that the traditional DBSCAN(Density-Based Spatial Clustering of Applications with Noise)algorithm only considers the spatial attribute on the problem of extracting the stay point and ignores the time attribute,the trajectory is further divided after clustering.The experiment is carried out by using the real data collected by our system and the results show that the proposed algorithm has a good effect.Finally,we design a set of position prediction scheme for the system.Considering the prediction accuracy of the traditional first order Markov predictor is not high and higher order Markov predictor has the problem of spatial expansion,we designed a hybrid multi-step Markov model.And the genetic algorithm is used to optimize each step of the weight coefficient.Through theoretical analysis and experimental evaluation,it is proved that the position prediction scheme designed in this thesis can balance the space-time complexity and accuracy.