Empirical Analysis Of GDP In Jiangxi Province Based On A Non-parametric Combination Forecasting Model

In recent years,the economy of Jiangxi Province has maintained a stable and rising development trend,but due to the impact of the new crown pneumonia epidemic,the economic development of Jiangxi Province will inevitably be affected.GDP as an important macroeconomic indicator,in the current world economy presents a complex and uncertain background,scientific and accurate forecasting of GDP to improve the level of scientific decision-making of the government has certain practical significance.In order to achieve accurate prediction of GDP in Jiangxi Province,this paper firstly selects five single models,ARIMA model,ARIMAX model,grey GM(1,1)model,BP neural network model and non-parametric autoregressive model for the prediction of GDP in Jiangxi Province;secondly,uses four evaluation criteria,namely,mean absolute percentage error,standard deviation of prediction error,Thiel’s inequality coefficient and prediction accuracy,to evaluate the Then,the ARIMAX model and the BP neural network model were selected to build four combined forecasting models based on the equal weight method,the simple weighted average method,the inverse of the squared error method and the inverse of the mean squared error variance method according to the forecasting characteristics of each single model.Since the weight coefficients corresponding to each individual model in these four combined models are fixed,it is impossible to change the magnitude of the weight coefficients according to the prediction effect at each moment,which leads to the fact that the established combined models cannot make maximum use of the effective information of the data;finally,according to the data characteristics of GDP and combining the prediction advantages of each individual prediction model,this paper introduces a nonparametric autoregressive in the ARIMA model method proposes a nonparametric combined forecasting model method based on residual correction.By forecasting the GDP of Jiangxi Province and comparing the forecasting effects of each model under the established evaluation index system,the results of the empirical analysis show that the paper establishes a non-parametric combination forecasting model based on residual correction.The results of the empirical analysis show that the combined forecast model based on residual correction established in this paper has higher forecasting accuracy than the four combined models formed by giving different weights to each individual model and the individual models before combining,and also outperforms other models under the remaining evaluation indexes.According to the results of the non-parametric combination model,the GDP of Jiangxi Province will grow at an average annual rate of 7.01% during the 14 th Five-Year Plan period,and the total GDP will be RMB 282.8026 billion,RMB 308.8440 billion,RMB 326.1872 billion,RMB 347.7033 billion and RMB 370.4525 billion.