Statistical Analysis And Modeling Of Domestic Tourism Data

With the rapid development of tourism recently,it makes great contribution to the development of our country's economy.It has become one of the main industry of the national economy and the main component of the service sector.It makes an positive effect on the adjustment of industrial structure,promoting the consumption etc.Tourism has greatly promoted the development of economy,so it worths to be aimed at the data of domestic tourism to the statistical analysis and modeling.In order to analyze the primary factors influencing the revenue of domestic tourism,we use the partial linear model in this paper.As a semiparametric model,the partial linear model is widely used to deal with practical problems,not only has the metirs of linear model which is facilitate to explain,but also has a better flexibility.Firstly,we introduce the related achievements of the partially linear models,and we introduce the theories of the two methods of the parameter estimation,such as the method of kernel estimation and the method of partial smooth spline estimation.Then We use the methods to make the numerical simulation,and we compare the merits of the two methods of estimation which under different sample sizes or the different variance of the error.Finally,We carry out an actual example,it selects seven indicators which affect the revenue of domestic tourism.The dimensions of variables are reduced by using the principal component analysis.Then we use the the partial smooth spline estimation to estimate the unknown parameter of the partially linear model,and the domestic tourism income in 2015 is forecasted according to the model.For contrast,a linear model is also constructed with the data.According to the comparison of the relative errors,I find that the partial model has a better performance,it worths to be applied,and the conclusions and suggestions are finally given.