User Information Mining Technology Research Based On Satellite Positioning Data

With the progress of the times and the great improvement in people's living standards,more and more people are buying cars.On the one hand,with the development of various positioning technologies and the popularity of mobile devices,it has become easier and easier to obtain people's positioning information and convert it into spatio-temporal data.On the other hand,in order to maintain the rights of car service companies and consumers,ensure people's vehicle safety,and facilitate efficient dispatch of vehicles,many vehicles are equipped with vehicle positioning equipments,which continuously send positioning data to specific control centers.These data contain a large amount of information,provide rich data resources for analyzing the user's behavior habits and other information,and are beneficial to the realization of various location-based services,which is of great significance for risk control,urban construction and route planning.This thesis analyzes the user information mining technology based on satellite positioning data,and conducts in-depth research on resident point discrimination,similarity measure,track point clustering and user information mining and visualization.Based on the real positioning data,the user's resident information,frequent items and important point information are mined.The specific innovations are as follows:(1)Residing point discrimination.Through the research on the mining process of location data,aiming at the discontinuity and incompleteness of location data,the algorithm of traditional dynamic resident point discrimination is improved.In this thesis,a SMCT algorithm is proposed,and its predefined parameters are analyzed in detail.And this thesis verified that SMCT algorithm has better performance and effect than traditional dynamic resident point discrimination method by accuracy,precision and F value,and proved that the user residence points extracted by it have the characteristics of no residence bias and stronger applicability.(2)Similarity measure.This thesis studies the representation of time,proposes a time point similarity measurement method based on DT distance,and combines it with clustering algorithm to mine users' time information.(3)Track point clustering.By studying the trajectory point clustering algorithm,on the basis of DBSCAN clustering algorithm,the author proposes TP-DBSCAN algorithm.By introducing the weighting coefficient and density threshold adaptive,the sensitivity of the algorithm to the parameters is reduced,and the problem that the neighborhood radius of the track point is inversely proportional to the value density is solved.And the contour coefficient is compared with other algorithms to verify that the TP-DBSCAN algorithm has good clustering effect and stronger robustness.(4)User information mining and visualization.Based on the real satellite positioning data mining,user's residence information,frequent items and important information are mined in this thesis.Combined with the high-tech map,the user's important location is semantically defined,and the user information is visualized.