Font Size: a A A

Prediction Of Local Fiscal Revenue Based On The Gray Markov And RBF Neural Network Model

Posted on:2022-07-27Degree:MasterType:Thesis
Country:ChinaCandidate:P Y LiuFull Text:PDF
GTID:2480306737953329Subject:Applied Statistics
Abstract/Summary:PDF Full Text Request
Under the current tax sharing system in China,local fiscal revenue,reflecting the level of national economy of a local,is an important part of the national fiscal revenue.It is the basis of the government's macro-control.The economy Hunan Province is stable in the middle and long term.However,the pressure of economic development increases and the scale of tax reduction and fee reduction policy is also expanding,which make the growth power of fiscal revenue weaker than before.Therefore,Scientific and reasonable forecast of local fiscal revenue not only can make the revenue forecast and expenditure arrangement in the budget no longer random and blind but also can promote the correct handling of the relationship between finance and economy.This paper takes the fiscal revenue and its influencing factors of Hunan Province from 1994 to 2019 as the research object and forecasts the fiscal revenue from 2020 to 2022 by using variable selection method and forecasting method.First,according to the literature,22 factors influencing the fiscal revenue of Hunan Province were selected.Eight explanatory variables closely related to Hunan financial income were selected by using the Variable Importance Measures of Random Forest.These variables include the first industry,total exports,total population,the number of employees and so on.Secondly,the short-term prediction of these 8 explanatory variables is carried out by using the gray Markov model.Finally,by using back propagation(BP)and radial basis function(RBF)neural network,the forecast value of Hunan fiscal revenue from 2020 to 2022 is obtained.In the network,the eight explanatory variables selected are used as the input of the network,and the fiscal revenue of Hunan Province is used as the output of the network.The two models are combined by standard deviation method.Finally,the rationality of the prediction results is expounded from the two aspects of the correlation between the fiscal revenue and its main influencing factors and the growth rate of the fiscal revenue in Hunan Province.
Keywords/Search Tags:the fiscal revenue of Hunan Province, random forest, gray markov model, RBF neural network
PDF Full Text Request
Related items