Research On Ningxia Industrial Economic Prosperity Index And Forecasting Model

With the rapid development of high-tech industry and strategic new industries,our country has become a veritable big industrial country.Because the industrial data has the characteristic of periodicity and weak elasticity,it is necessary to analyze and forecast the development trend of industrial economy.Therefore,many scholars apply the industrial prosperity index to analyze the development of industrial economy.However,the research on the prosperity index is mainly concentrated in the macroeconomic field at the national level,and there are relatively few researches on provincial industrial prosperity index and the accurate prediction of industry-related indicators.In this context,this paper studies Ningxia industrial economy based on principal component analysis method and predicts the benchmark indicator of Ningxia industrial prosperity index based on forecast model.The following contents are mainly studied:(1)Considering the characteristics of Ningxia’s industrial structure,the leading index,coincidence index and lagging prosperity index of Ningxia’s industry were compiled based on principal component analysis.By observing the complete fluctuation cycle,it is found that the fluctuation cycle of the prosperity index is about seventeen months;By observing the fluctuation of peaks or troughs of different prosperity indexes,the leading index leads the coincidence index about six months,and the coincidence index is also about six months ahead of the lagging index.(2)Based on Granger causality test,Impulse response function analysis and Variance decomposition,the rationality and dynamic interaction of Ningxia industrial prosperity index are explored.The research shows that the leading index is the Granger reason of the coincidence index,it is that the leading index can predict the cyclical fluctuation of Ningxia industrial economy;In the short time,both the leading index and the coincidence index promote and influence each other,and the coincidence index has a great influence on the leading index.The coincidence index and the lagging index have the same change trend,and the two indexes show fluctuating changes,and the coincidence index has a great influence on the lagging index.(3)Based on the hybrid method(ARIMA-Bayes-LS TM)which combines autoregressive integrated moving average(ARIMA)model and Bayesian optimization for long short term memory(LSTM)model to forecast the benchmark index of Ningxia industrial prosperity index.In this paper,the ARIMA model is used to forecast the linear part,and the residual only contains nonlinear relations.Then,the Bayesian optimized LSTM model is used to predict the residual,which reduces the uncertainty of the model and parameters.The result shows that the combined forecasting model of ARIMA-Bayes-LSTM has the best prediction performance and reduces the forecasting risk of a single model.The work done in this paper provides scientific data support for the development of Ningxia industrial economy,and has certain guiding significance for monitoring the trend of industrial economy and looking forward to the economic operation situation.It can provide the basis for local policy control and monitor the healthy development of Ningxia’s industrial economy.