Neural networks are remarkable tools for extracting data changing patterns with the existence of noise distraction. Thus neural networks are widely used for forecasting problems, for instance weather forecasts and airport passenger volume forecasts. Airport passenger volume forecasting is extremely useful to predict the passenger demand in the future. Based on the passenger movement forecast, the airport managing group can adjust the airport service, resources and facilities to meet the passenger demand while improving the airport efficiency and resource usage. Airport passenger forecasts can be classified into short-term forecasts, medium-term forecasts and long-term forecasts. Long term passenger movement forecasting plays an important role for airport expansion plans, including new airport terminal construction.  Here we focus on the medium-term forecasts. If the forecasting model is accurate and has a high confident level, the airport group can make development plans based on the forecasting results in order to cater the future passenger demand. In this case, we use the passenger data of Changi Airport for the past 20 years and data from the Singapore Statistics to train our neural network model and make a prediction for the future passenger demand.

neural network, passenger volume forecasting, MATLAB simulation, statistical analysis, Changi Airport.