Study On The Prediction Of Construction Costs

In market economy system, construction cost prediction,which is regarded as an essential component of construction cost administration system, not merely is the key that owners and contractors maintain the competitive advantage and obtain the profits, but also is the basis on which the government carries on the macroeconomic regulation and control and other correlation benefit main bodies in the construction market make investment decision and developmental strategy. Construction cost prediction is studied thoroughly abroad,but it is still at the start stage and the quantitative research articles on it are few in our country as a result of the historical problems. For this reason,this thesis adopts correlation analysis method and neural network theory to study domestic construction cost prediction on the base of the qualitative analysis of the relevant factors which affact it. The main research work and results are as follows:First of all,this thesis analyses all the economic statistic indexes related to construction cost ,which is involved in 《China Statistics Yearbook》 of 1992-2004 , from angles of economics and management and 14 aspects ,and elicits 44 marcro or micro interrelated fators of construction cost. These factors reflects various aspects which influences construction cost comprehensively.Secondly, this thesis carries on quantitative analysis to the 44 marcro or micro correlated factors with correlation analysis method and elicits 15 high correlated factors.Because the historical data of the factors is the foundation of predicting model and too many factors do not always lead a retional result.Finally , this thesis sets up BP nerve network predicting model on the basis of historical data of the 15 high correlated factors. The precision within the sample meets the requirement.The predicting model is empolyed to forecast the construction cost of the future years.The result of it shows that the construction cost in the next 5 years assume the...