Cost indexes are a method of comparing cost changes from period to period for a fixed quantity of goods or services. Cost indexes are useful in construction as a method of forecasting the cost of similar designs to future periods without going through detailed costing. Two back propagation neural network models have been developed to predict changes in the ENR Magazine construction cost index. The models make predictions for one month and six months in the future. The models were developed using the Neuroshell computer program. The training set consists of historical economic data for the period from 1967 to 1991. The three layer back propagation models did not make accurate predictions. The neural network models were compared to forecasts made by exponential smoothing and linear regression. Exponential smoothing produced forecasts with the lowest error. The results of this model development indicated that predicting changes in construction costs is a complex problem. Changes in construction cost indexes could not be accurately predicted. Additional work is required using different neural network models, such as back propagation models with more hidden layers. More work is also needed to find good economic indicators of changes in construction prices.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £34 +P&P)