Tax Revenue Prediction Of Heilongjiang Province Based On Combined Model

Tax is the basis of a country's viability and development. And tax revenue is the significant index to measure a country's financial resources and the government functions. Under the market economy condition, in case of tax rooted in economy, a suitable tax planning measure is a direct connection with tax and economic development, and the scientific tax forecasting is premise condition of tax planning. On the basis of review on domestic and overseas publications about tax forecasting, this issue focuses on the relationship between tax and economy and builds Heilongjiang Province Tax Forecasting Model. Firstly to formulate tax plan according to the local realities.At first, basis on the analysis of research background of tax forecasting and amount of review on domestic and overseas publications about tax forecasting to discuss the research purpose and significance of this issue. The author chooses economic index with analysis of the relationship between the amount of tax revenue and economy, and builds Stepwise Regression Tax Forecasting Model in order to obtain forecasting results with empirical test. Based on the previous chapter to build BP Neural Network Forecasting Model, and obtains forecasting results with extracting and testing historical data. The author concludes that BP Neural Network Forecasting Model has higher prediction accuracy from comparing predictive value with actual value.Based on minimum variance estimate to weight distribution, and combines Stepwise Regression and BP Neural Network Forecasting Mode. And the author concludes that the efficiency of combination forecasting is more accurate and steady than that of any single model. Combination Forecasting Model improves the efficiency of Forecasting Model.