Research On Hot Spots Recommendation Of Urban Traffic Demand Based On Trajectory Data

The accelerating of urbanization process and the improvement of people’s living standard have led to the rapid increase of private cars as well as increased pressure of urban road traffic.Therefore,it is urgently needed to guide the urban road planning and improve the level of urban management through the large amounts of analyzed traffic data,and find the laws of urban operation in the traffic data.Meanwhile,with the rapid development of wireless communication technology and intelligent mobile terminal,the acquisition of trajectory data of moving objects becomes more convenient.The data of taxi trajectory is easy to be collected,widely distributed and with large quantity,which make the data mining of GPS data of taxi trajectory the research hotspot in recent years.Currently,problems existed in the taxi industry are high no-load rate and difficulties of hailing a taxi.Therefore,it is of vital research significance to provide recommended service for taxi drivers.In this paper,through the study of large amounts of taxi GPS data,the taxi stay points are analyzed,the hidden rules of trajectory data are discovered by the way of data mining technology,the taxi hot spots area and recommended service are conducted in-depth research.The purpose of this paper is to minimize taxis’ no-load rate,reduce urban traffic pollution,ease the traffic pressure and provide valuable references for the operation and management of the taxi.Firstly,the taxi GPS data and road network data are preprocessed for later data analysis and processing.Road network are edited secondly through ArcGIS platform,which include topology processor of road network data,perfection of the property field,the establishment and verification of road network data set.Map-Matching for Low-Sampling-Rate GPS Trajectories is adopted,and with the constraints of road network topology and speed,the location information of the vehicle is compared with electronic map network information to determine the real situation of vehicles on the road network.Secondly,the laws that the taxi stay points change over time are obtained by statistical analysis,a semi-supervised neighbor propagation algorithm based on particle swarm(PSO-SAP)is proposed for discovering taxi candidate hot spots,PSO-SAP is integrated on ArcGIS platform,trajectory data is analyzed by spatial-temporal analysis,Then,taking advantage of the ArcGIS platform to achieve hot spots area display and figure out taxis’ distribution of hot spot area at different times.Finally,recommendation algorithm based on trust degree of the taxi drivers is put forward to recommend hot spots within the area of large passenger probability,the tests shows that recommendation result enjoys higher accuracy.