Research Of Regional Fixed Asset Investment Forecasting

Investment in fixed assets has significant effects on China's macroeconomic operation,especially in promoting economic growth,optimizing the al ocation of resources,improving people's living and expanding employment,etc.Therefore,it is necessary to make an accurate forecasting of investment in fixed assets.Fixed asset investment tends to be influenced by many uncertain factors,which are complicated and difficult to predict by mathematical statistical models.This thesis uses the grey neural network model,support vector machine model and ARIMA model to forecast the fixed asset investment in TianJin respectively,and has carried out comparative analysis.On the basis of studying current fixed asset investment forecasting at home and abroad,this thesis further analyzes the theory of forecasting model and explores the forecasting mechanism.In this thesis,the forecasting analysis of support vector machine model is realized by MATLAB,and the forecasting analysis of the grey neural network model and ARIMA model are realized by R language programming.Through empirical analysis,the support vector machine model and ARIMA model have higher forecasting accuracy,and the forecasting accuracy of grey neural network model is lower by contrast.On the forecasting trend effect,the predicted value of the support vector machine model and ARIMA model fluctuate near the actual value,the predicted value of grey neural network model was slightly lower than the actual value.Meanwhile,the predicted curve of grey neural network model was very similar to the actual curve,suitable for long-term trend forecasting;On the forecasting randomness,ARIMA model shows the shape of a broken line compared to the other two forecasting models,embodying the characteristics of random time series.So ARIMA model is more suitable for short-term forecasting.Finally,this thesis summarized and prospected the research.This thesis has carried out effective forecasting of fixed assets investment in TianJin by using three kinds of forecasting model,and has verified the rationality of the forecasting model and explored the advantages and disadvantages of each model.At the same time,this thesis puts forward some problems to be explored and improved in the future.At last,some reasonable suggestions for the fixed assets investment are put forward for TianJin.