Research On Domestic Tourism Market Forecasting Based On ARIMA And SVM Combination Model

With the growth of China’s economy and national income,the tourism industry has become popular increasingly and shows great vitality in the political,economic,social,cultural,ecological and other fields since the reform and opening up.At present,China’s domestic tourism has become the tourism market of the fastest rate,the largest number,the most potentiality in the world,which has exceeded inbound and outbound tourism in the development scale and occupied an important position extremely in the economic development.Therefore,how to grasp and predict the development trend of the domestic tourism market in the future,make the right decision for the tourism enterprise and the manager becomes the realistic problem that the government must solve.Firstly,the background of the domestic tourism market are analyzed,discusses the research contents,methods,and significances.Then the thesis adopts the grey correlation degree to analyze the influence factors data of domestic tourism market in 1995-2015,and judge the accuracy of each model that predict the number of domestic tourists by the mean absolute percentage error(MAPE)size.On the basis of it,traffic and city residents per capita disposable income are predicted which is used to predict domestic tourists.Lastly,the thesis puts forward some suggestions according to the continued growths of domestic tourists bring some problems(traffic congestion,tourism safety,the decline of tourism service quality and market disorder)in the future.The main conclusions of this thesis are as follows:(1)The thesis gives a correlation analysis of influence factors of domestic market(traffic(total mileage of highway and railway),tourism product prices(CPI),tourism environment(domestic tourism revenue)and city residents per capita disposable income)with gray correlation analysis method,the results show that traffic is the most important,followed by city residents per capita disposable income.(2)To predict domestic tourists by comparing with differential autoregressive moving average(ARIMA),support vector machine(single factor SVM and multi factor SVM)combined model and considering the characteristics of linear and nonlinear of domestic tourists,the results show that the combined forecasting model have higher accuracy,stronger generalization ability.(3)The number of domestic tourists and transportation(total mileage of highway and railway),tourism product prices(CPI),tourism environment(domestic tourism revenue)and city residents per capita disposable income have a strong positive correlation,especially in the early stage of economic development,economic development has a strong dependence on domestic tourists and uses combined model to predict the two most important influencing factors of domestic tourists(traffic and disposable income)which has stronger generalization ability and more accuracy,then through the results of the three polynomial regression to predict,find that the traffic to predict the number of domestic tourists has more reliable.(4)The number of domestic tourists increased relatively stable in 1995-2004,but the increased trend of tourists accelerated in 2005-2015.From the prediction results,the number of domestic tourists will be increased,tourism market size will be further expanded,Per capita travel rate will be 9.6 times and tourism consumption will be greatly improved in next ten years,which are coincided with the strong support of the state on the tourism and economic structure adjustment and transformation.