The Domestic Tourism Demand Forecasting Based On The Combination Model Of Linear And Nonlinear

In recent years,with the amazing speed of tourismdevelopment,tourism has becoming one of the industrial which has themost powerful develop ment momentum.It promotes economic developmentstrongl y, more and more countries or regions put it as a pillar industry anddevelop tourism vigorousl y and hop e it can drive the development of thewhole social economy.Therefore,the tourism demand forecast c an providea basis for the tourism authorities to make the development policy andstrategic planning,so the tourism authorities can guide the optimalallocation of market resources,and the tourism enterprises can get thereference for their own development.Currentl y,the models of domestic tourism demand anal ysis andforecasting have the general linear regression models、 the time seriesmodel、the gray model、the artificial neural network model and so on.Thesemodels can be used for domestic tourism demand for ecasting,but there isno uniform conclusion to use what kind of model in specific cases.Thispaper uses five commonl y used regression models、 gray system theory、BP neural network model and the combination model to anal ysis andforecast the domestic tourism demand.Firstl y,we anal yze the related factors of tourism and use the fivecommonl y used regression models to forecast and anal yze the domestictourism demand,and we compare the advantages and disadvantages ofdifferent models;Secondl y,we use the gray corr elation degree to anal yzethe influencing factors which have different influence degrees to domestictourism,and use the GM(1,1) model to forecast the domestictourism;Thirdl y,we use the classic BP neural network model to anal yze thedomestic tourism deman d,and use the BP neural network model to correctthe GM(1,1) model,thus,the forecasting accuracy of corrected GM(1,1)model can be improved;Finall y,we use the BP neural network to combinethe corrected GM(1,1) model and the LASSO regression model,and usingthe combination model to forecast the number of tourists,the result showsthat the combination model can include more information and has thehigher prediction accuracy.