Font Size: a A A

Design And Implementation Of Accompany Analysis And Community Detection System Based On Trajectory Data

Posted on:2019-12-19Degree:MasterType:Thesis
Country:ChinaCandidate:C C ZhangFull Text:PDF
GTID:2428330566497307Subject:Software engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of mobile Internet technology,smart mobile terminals have already had a profound impact on people's lifestyle.In this context,this system is based on researching the spatio-temporal data of users in the smart mobile terminal network lo g,acquiring spatio-temporal data therein,extracting location information and constructing user spatio-temporal data according to time,based on a large number of individual user spatio-temporal data,mining The characteristics of the user's data in the s pace-time,the characteristics of the user's spatio-temporal data in accordance with the adjoint pattern are analyzed,and the community relationship is further analyzed with the accompanying results.The system proposes the analysis of group adjoint mode and community discovery based on user trajectory data,and describes accompanying information from the start time,the end time,the number of participants,the participating users,and the accompanying state.Design and implement system clustering operators and discovery algorithms based on the Map Reduce parallel processing computing model and the Spark big data processing engine.Time-slicing techniques and linear interpolation are used to model user trajectories,and a standardized and complete user tra jectory data is constructed.A DBSCAN clustering algorithm for user location data that is applied to a single time slice is designed to obtain clusters containing user location data and basic information that meet the density requirements.Designed and imp lemented a crowd discovery algorithm that meets the characteristics of scale,density,duration,and geometric stability suitable for parallel computing environments.A contiguous verification algorithm for discovering user-accompanying information on the basis of discovered people is designed and implemented.On the basis of obtaining the concomitant results,the community discovery algorithm based on Louvain algorithm is further designed and implemented.The display interface was designed to display the results in the form of Web.A simple and userfriendly interface was realized by accessing the plug-ins of GIS service providers.Redis was used as a real-time presentation database to utilize its fast read/write characteristics to support the real-time presentation of aggregated results.My SQL,as a historical storage database,implements the storage of historical analysis results.The accompanying results are stored in the graph database Neo4 j in a certain model.Finally,this paper designs the test method of the system,including functional testing and performance testing.During the testing process,the program is optimized to achieve industrial use conditions.At present,the system is in normal operation.
Keywords/Search Tags:Trajectory data, Accompany model, DBSCAN, community discovery, Louvain
PDF Full Text Request
Related items