| Time series is exist in various fields,which has characterizes of high dimensional,dynamic,complex,high noise and has a large amount of data.Model is an expression of time series.The fitting model will get a bad result,if you just use a set of basis functions on the whole domain,when the data of time series has a lot of significant local features.The piecewise polynomial regression fitting model is one of the effective methods to solve such problem.If you want to use the piecewise polynomial regression to fit the time series,the connection on the border is particularly critical.This article builds three fitting models,which mainly bases on the population of Shanghai outbound tourism.The first is using seasonal model to fitting the population of Shanghai outbound tourists.The second is combine piecewise polynomial regression with cubic spline interpolation to establish a global continuous piecewise polynomial regression model.The result shows that the second model is better than first one.The third is establishing a global smooth piecewise polynomial regression model.The contribution:(1)Based on the piecewise polynomial regression and cubic spline interpolation to establish a global continuous piecewise polynomial regression model.It is applied to fit the population of Shanghai outbound tourism at first time.The fitting result of the global continuous piecewise polynomial regression model is better than seasonal model.(2)Building a global smooth piecewise polynomial regression model,which based on the global continuous piecewise polynomial regression model.This model is used to fit the population of Shanghai outbound tourism at first time.It solves the rough connection of piecewise polynomial regression model. |
