The Seasonal Adjustment And Application Of Internet Search Weekly Unemployment Rate Data Based On MoveReg Method

The rapid development of network information brings the background of big data.In this paper,the weekly internet search data,as an assistive tools of unemployment rate detection,is employed to overcome the disadvantages of low frequency and insufficient statistical objects in China's unemployment statistics.Firstly,the MoveReg method and structural time series model method are introduced for the weekly seasonal adjustment,and its feasibility is analyzed in detail.Then,the Letterman interpolation method and three-step word technical selection method are used to accomplish the internet search keyword selection.As such,a seasonal adjustment of internet search weekly unemployment rate data,along with the comprehensive analysis on the results,is done.Finally,the principal component analysis synthesis method is used to synthesize the seasonally adjusted data into an index,then,introducing the output gap,the lagging term of the unemployment rate and the lagging term of the output gap to establish the traditional linear Austrian affirmative law model,and then combines the smooth transformation regression model to construct the nonlinear Austrian affirmative law model.The existence and nonlinear characteristics of affirmative law in China are investigated.The results show that the use of MoveReg method has a good effect on seasonal adjustment of internet search weekly unemployment rate data,and there is a long-term equilibrium relationship between internet search unemployment rate data and output gap.Although there is no obvious linear relationship between the two,however,the influence in the nonlinear part is significant.It can be considered that the proposed internet search data based weekly seasonal adjustment would accurately reflect the behavior of unemployment rate in time,which also fill the research gap in China's weekly seasonal adjustment.