Citation

Yabo X, Tong W, Bingxue S, Yushu X. Procedia Eng. 2012; 43: 132-136.

Copyright

(Copyright © 2012, Elsevier Publishing)

DOI

10.1016/j.proeng.2012.08.023

PMID

unavailable

Abstract

The forecast of production safety situation is a complicated non-linear problem. The developmental change possesses no obvious trend of change over time and random fluctuation. Taking construction industry data as an example, four forecast models are adopted separately, namely the back-propagation neural network model, the moving average model, the exponential smoothing model and the combination model. Estimated results show that the combination forecast model can overcome the shortcomings of the single prediction model, and solve the forecast difficulties caused by random changes of the number of safety indicators of system status. The combination model is feasible for the forecast of the construction industry production safety situation. In fact, the production safe situation is affected by time, policies and other related factors. Decision makers should dialectically use the forecast result in the actual application and may carry on the weight adjustment to the forecast value.