Neural Network Forecasting of the Production Level of Chinese Construction Industry
Much more efforts have been devoted over the past several decades to the development and improvement of time series forecasting models. In this paper, the authors wish to determine whether the forecasting performance of variable under study can be improved using neural network models. Among best 10 retained networks, a MLP 3-layer network: 1:1-31-1:1 is selected as the ANN model with the minimum RMSE. The best network found has excellent performance (regression ratio 0.003343, correlation 1.000000, and error 0.000147). The optimum empirical neural network obtained has a satisfactory degree of statistical validity (low approximation errors i.e. RMSE=0.0078 and MAE=0.00518). The performance of the model is evaluated by comparing with ARIMA model. The root mean squared forecast error of the best neural network model is 49 per cent lower than the ARIMA model counterpart. It shows that the neural network yield significant forecast improvements. The gains in forecast accuracy seem to originate from the ability of neural networks to capture asymmetric relationships. As the Chinese construction industry shares in GDP considerably, it has important and supportive role in its national economy. This paper is also acknowledged the situation of the construction industry and its implication to overall economy. The empirical results show that the trend of production level of CI is increasing steadily implies the strong potential for future growth.
time series forecasting GDP production level of CI ANN model ARIMA model
DilliR. Aryal WANG Yaowu
School of Management, Harbin Institute of Technology. Harbin,Heilongjiang.150001 School of Management, Harbin Institute of Technology, Harbin, Heilongjiang. 150001
国际会议
2003 International Conference on Construction & Real Estate Management(2003 建设与房地产管理国际会议)
哈尔滨
英文
319-326
2003-11-20（万方平台首次上网日期，不代表论文的发表时间）