Passenger Flow Distribution Of Rail Transit Based On Navie Bayes Classifier

With the continuous expansion of scale of modern city and the population gradually concentrated on large cities,in order to solve the increasingly serious traffic problem,rail transit has become a priority choice for metropolises to handle traffic congestion and travel difficulties due to its characteristics of large capacity,punctuality,efficiency,safety and environmental protection.During the planning of rail transit,the process of transferring is usually under the same station,which means that passenger can transfer from one line to another without getting out of the station.It does bring a great deal of convenience to passenger’s travel,and however,the exact route in rail transit networks chosen by individual is under unknown since they only need to swip their IC card when going in and out of station.On the one hand,due to the operation of capital,the PPP(Public-Private Partnership)mode is adopted to construction of rail transit,which means the construction of public infrastructure is invested by government and social capital together and earnings generated from rail transit’s tickets is also shared by them.On the basis of network for rail transit,if passenger flow distribution of rail transit can not be mastered accurately,the cooperative relationship between enterprises and government will be overshadowed by it.On the other hand,from operators of rail transit point of view,knowing passenger’s route choice and even passenger flow distribution of rail transit is of great significance for improving service efficiency and improving service level.Most of the existing methods are based on the impedance function of the path to obtain the probability of the path,and spending time on path,while the main body of one trip,individual,is ignored.For multipath problems,the factors that affect the passenger’s choice of path and even the passengers themselves are the core of the problem.Based on the premise of network for rail transit,the paper considers three key factors influencing the passenger route choice: travel time,travel distance and transfer times,with an individual as the research object,and proposes a Naive Bayes Classifier model based on comparing the value of posterior probability of each path to determine the exact route chosen by passengers.First,this paper analyzes the factors that influence passenger’s choice of path,and divides them into deterministic factors and uncertain factors,and then analyzes the distribution rule of travel time as well as the composition of travel time,so as to lay the foundation for calculating the value of travel time.Second,the definition of effective path is proposed.By combining path impedance function and breadth first search algorithm,the reachability and impedance of path are considered together,and effective path set can be generated between OD.Third,according to the principle of Naive Bayes Classifier,a passenger flow assignment model,regarding travel time,travel distance and transfer times as attributes,and single passenger as the research object,is proposed.Finally,taking the rail transit system in Guangzhou as an example,the distribution of passenger flow,the distribution of passenger flow at peak hours and the distribution of passenger flow on the line network are analyzed,and the rationality of the calculation results is analyzed too.