Research On Forecasting Algorithm Of Inbound Tourism Demand In China

In recent years,the tourism industry in China has developed rapidly and becomes an important motivation for economy.Among them,as one of the key indicators to measure the level of tourism development in our country,the role of inbound tourism in China's transition from a large tourism country to a stronger one has become increasingly prominent.The existing inbound tourism data modeling and predictive analysis will provide a scientific decision-making basis for the further development of inbound tourism,which is related to the work of the tourism management departments and industry.Based on the data released by the National Tourism Department,this paper analyze the situation of inbound tourism from the perspective of supply and demand in China.Especially predict the number of the inbound tourism passengers with the data from 2000 to 2016 by modeling the Seasonal Autoregressive Integrated Moving Average model,the Holt-Winters algorithm and the Multilayer Perceptron model,and then compare the prediction results of the three models.The model with the lowest average absolute percentage error is selected to predict the number of the inbound tourism passengers in 2017,which is consistent with the actual result.In addition,when building a Multilayer Perceptron model,a data preprocessing method using lateral prediction in one-dimensional time series is used to improve the prediction accuracy of the model.