Early-warning Technology Of Railway Passenger Tickets Sale Based On The Classification Learning

Compared to civil aviation,the competition of railway passenger transport in the transport capacity among passengers from different OD is more intense.And the contradiction is particularly acute in Chinese railway passenger transport,because of its operating mileage is generally longer and its railway capacity tensions remains to be solved.To resolve this contradiction,Chinese railway passenger transportation department has gradually formed a series of ticket sale organization strategy system which is used to adjust the railway capacity to meet the need of passengers from different OD.And the basic principle of ticket sale organization strategy is giving priority to long-distance passenger,trying to treat the short-distance passenger fairly.However,as for forecasting of railway passenger flow before the booking period is difficult to meet the ideal accuracy.Therefore,it is necessary to adjust ticket sale organization strategy to match the market demand immediately by making an early-warning on the train that may have an irrational strategy during the ticket booking period.For this purpose,this paper tries to propose an early-warning method for railway passenger ticket sales using classification algorithm,and this early-warning method is trying to overcome the problem of inefficiency that relies mainly on managers’ manual judgment at present,the intractable problem to adapt the train increase for the past few years and deficiencies that operational management requirements are increasingly refined.This paper gives a detailed introduction to the business of tickets organization dynamic adjustment and early-warning,and then gives a clear definition to the demand for the ticket sales early-warning in booking period.With the analysis of train operation history data in the Railway passenger transportation marketing auxiliary decision system,the paper proposes some property variables describing the characteristics of the train as well as the characteristics of the passenger ticket sales process,and then summarizes five type possible sale results.On this basis,On this basis,construct the passenger ticket sales early-warning model that combined with the advantages of the bayesian classification in uncertain reasoning.then uses this model to evaluate the process of passenger train ticket sale.Furthermore,based on the evaluation result,put forward a method of early-warning for trains that may need to be adjusted.At last,the paper applies this early-warning method on train with the data of train history operations that from Beijing shanghai high speed railway,and it demonstrates the validity and the applicability of this method,At the same time,we analyze the accuracy of this early-warning method from different angles,such as travel date,the time before departure,passenger attendance rate,etc.