Research On The Recommendation Method And System Implementation Of Taxi Pick-up Hotspots Area

Taxi is an important mode of transportation for residents.As the demand for taxis increases,the taxis and passengers often have unbalanced supply and demand during the taxi operation,resulting in passengers taking taxi difficult and taxi empty driving.With the popularization of GPS,the taxi GPS equipment produces a large amount of trajectory data.This paper studies the taxi trajectory data,minings and recommends taxi pick-up hotspots area,which can provide guidance for taxi drivers and solve the problems of passenger taxi difficulty and empty taxi driving.(1)In the area of passenger hotspots area mining,the existing research on taxi pick-up hotspots area mining mainly uses density clustering method.Aiming at the problems of insufficient clustering precision and high time complexity of existing density clustering algorithms,this paper presents a taxi pick-up hotspots area mining method based on A*routing algorithm and DBSCAN density clustering algorithm.A*algorithm is used to find the path and store the results in the adjacency table.A method of passenger point extraction is designed,which extracts passenger point by road section,in order to improve speed and accuracy.This paper uses Chengdu's floating car trajectory data and order data provided by DiDiChuXing for verification,compared with DBSCAN,this algorithm can mine more detailed taxi pick-up hotspots area and have better stability.(2)In the area of passenger hotspots area recommending,Because taxis are dynamic,the cost of going to different taxi pick-up hotspots area at different locations and times is different,so the passenger-carrying probability in different taxi pick-up hotspots area is also different.Traditional recommendation algorithms are not suitable for mobile point recommendation.Aiming at the problem of mobile point recommendation,this paper improves Item-based collaborative filtering.First,study the time distribution and spatial distribution,determine the peak time period,and select the peak time period for subsequent research.Then,through the Item-based collaborative filtering,combined with Gaussian attenuation,improve the recommendation accuracy.Finally,the recommendation of passenger hotspots area is completed with the recommended taxi pick-up hotspots area and the heat of the taxi pick-up hotspots area hot spot area.(3)In the area of system implementation.This paper designs and implements the taxi pick-up hotspots recommendation system.The system uses Django framework and redevelops based on ArcGIS API for JavaScript.The system is B/S structure.It can complete the functions of mining hot spots and recommending hot spots.It has practical value and provides the basis for mining results and recommending methods.