Research On Passenger Flow Prediction And Path Allocation Of Urban Rail Transit Based On AFC Data Mining

With the change of the development of the times,urban rail transit has gradually become the main artery of the city.However,China’s rail transit is facing huge safety risks,urban rail transit strives to find a reasonable way to predict the movement of short-term passenger flow,so as to take measures to deal with various situations that may occur during the peak of passenger flow.AFC(Automatic Fare Collection System),that is,the data provided by the urban rail transit automatic ticketing system,as the information that is easiest to reflect the direction of passenger flow,can provide sufficient data support for decision-making.However,due to the current lack of depth of AFC data mining,it has not been able to extract more valid information.This paper first sorts out the research status of urban rail transit passenger flow prediction and AFC data mining at home and abroad,and determines the idea of predicting short-term passenger flow based on AFC data information and using data mining and model establishment.Firstly,the processing method of AFC system data is introduced,and the stations are classified according to the existing passenger flow of urban rail,and the factors affecting passenger travel habits and judgment methods are given.Secondly,passengers are divided into passengers who have formed travel habits and passengers who have not formed travel habits,and data mining models and multiple Logit models are established to predict outbound stations.Then,a topological network is constructed for Chengdu rail transit,a "two-way search algorithm" is designed to generate an effective path set,and based on the improved Logit model and utility perception function,the matching probability of each effective path is calculated to obtain an effective path to achieve the purpose of predicting the passenger travel path.Finally,taking the real data of Chengdu Rail Transit in April as the data source for example calculation,the SQL language programming algorithm is borrowed to solve the outbound station prediction with the high-tech station as the O point,and the passenger path prediction of the Chadianzi to Yinghui Road.Compared with the real data of Chengdu Rail Transit in April,the model results can be obtained in line with the actual situation,and the verification speed is faster,which verifies the rationality and superiority of the model.The study shows that the algorithm of urban rail transit passenger flow prediction and path allocation based on AFC data mining can predict the passenger flow on the road network,and the algorithm model has high efficiency and rationality,which helps to guide the operation department to reasonably allocate transport capacity,provide a basis for the division of rail transit operating income,and also provide support and reference for the construction of new lines and the overall traffic efficiency improvement.