Neural network system for estimating light industrial buildings

Cost estimation for new projects is a lengthy and laborious process. It is a computational process by which estimators attempt to predict the cost of a future project. It usually has some degree of uncertainty involved because not all of the parameters and conditions for a project are known when the cost estimate is being prepared This limitation is even more pronounced when estimating for design-build projects. Many phenomena add to the noise in the final cost of a project which also make reliance on historical data difficult. Estimating methods, vary considerably depending upon various factors. Standard estimating models have been developed to facilitate the process by providing quick, "guesstimates" of the value to be used as a guideline Three different models were developed using linear regression, non-linear regression and artificial neural networks The neural network model was initially developed using the same parameters as inputs as for the regression models. Additional information was required to improve its output. Results obtained from statistical parametric models are compared with results obtained from neural network models.