关键词: 关键词：Holt-Winters模型, 客运量, 预测, 季节因素

Abstract: Abstract: Due to weather conditions, holidays, people's customs, population and economic growth and other factors, the volume of passenger traffic shows a cyclical growth trend. To make better plan for the passenger transport sector, in this paper, we use Holt-Winters model of the exponential smoothing method to decompose the time series data for seasonal and trends fluctuation, and use the monthly data of the volume of China's railway, road, river and airplane passenger traffic during 2002 to 2009 to fit. The results show that railways, airplane passenger traffic data have a clear linear trend and seasonal characteristics, and further we get the arrival time of peaks and troughs. Model has very good fitting results for them. The average percentage error (MAPE) are 5.536%，7.49%、6.070%和3.633%. On this basis, we conduct a scientific prediction for the volume of passenger traffic of each month in 2010.

Key words: Key words: Holt-Winter Model, Volume of Passenger Traffic, Forecast, Seasonal Factors