The Application Of Combination Forecast Model For The Air Passenger Traffic

The forecast of air passenger traffic is an important basis of organizing air passenger traffic.Before an important strategy made by air company,the forecast has been done.It can ensure to make scientific decision,workout feasible plan and layout and develop strategies.For the past few years,the passenger traffic of China has developed rapidly,the capacity of passenger and international status has improved significantly.The forecast of air passenger traffic is very necessary and urgertl.However,the air passenger traffic system is very complex,there are a lot of influential factors and the factors are uncertainty.Also,a unitary model cannot reflect the movement rule,so we will use combination model to forecast the capacity in order to improve precise.We will firstly introduce some theories of forecast,and analysis the conditions of their application briefly.Then,we use descriptive statistics to discusse the development status of air passenger traffic system in detail,we will see a phenomenon the time sequence of air fluctuates obviously by quarter.In this article,we used three models to forecast the capacity,they are quarterly index model,grayness model and back-propagation-neuralnetwork model.Through comparing three models,we conclude the advantage and disadvantage of them,and combined them to forecast time series.At the last,we will find the most suitable and accurate model according the performance appraising criterion,and use it to forecast the capacity of next year.This thesis applies combination forecasting to forecast the air passenger traffic,and build some combination model.According the empirical analysis,the results illustrates that the combination forecasting is very obvious.The research results can be directly applying to air passenger traffic,providing reference for aviation management,and offering some ideas for future researching.