Research On Recommendation Algorithm Of Taxi Stop Location Based On POI Regional Divide And Conquer

With the rapid development of Internet technology,coupled with the low efficiency of driver reception and the unreasonable allocation of resources led by the poor waiting location of taxis,the traditional taxi industry is facing the great competition of online car hailing.At the same time,the reasonable planning of transportation resources is also an important aspect of promoting urban development.In view of the above situation,we can combine spatial data query technology in complex network to solve such problems.Therefore,based on the actual social problems,this paper calculates the best parking location of taxis in different demand scenarios according to the user’s demand for taxis,and carries out in-depth research combined with the nature of the road network,dividing the whole road network into regions,a taxi parking location recommendation algorithm is designed.Firstly,this paper selects POI with large demand for taxis to construct undirected weighted graph.The probability transfer matrix is constructed according to the factors of taxi travel time,and the probability transfer matrix is used to guide the deep walk model to carry out truncated random walk in the road network graph.The high-dimensional sparse information contained in the nodes of the graph is mapped into low dimensional dense vectors,and the density clustering algorithm is applied to the region division in the vector representation space according to the similarity between the nodes.Then,a General Node Importance(GNI)algorithm for determining the importance of nodes is designed by selecting the appropriate indices according to the basic information of nodes in the road network.In order to solve the problem that the demand of taxis in different places is significantly different,a call distribution model of taxis on demand under different modes is established.Taking the number of taxi calls at the node as an attribute of POI under the background of this paper,based on the general algorithm,this paper puts forward an index of node importance —— node weighting coefficient,and designs a Dedicated Node Importance(DNI)algorithm based on POI.Finally,according to the ranking results of node importance under DNI algorithm,combining the leverage principle with the method of analogy polygon to obtain the center of gravity and considering the importance of nodes and the distance between nodes,an algorithm is designed to recommend the best taxi parking location,and the effectiveness and practicability of the proposed algorithm are verified by experiments based on real data sets.