摘要：Early stage cost estimate plays a significant role in any initial construction project decisions, despite the project scope has not yet been finalized and still very limited information regarding the detailed design is available during these early stages. This study aimed at developing an efficient model to estimate the cost of building construction projects at early stages using artificial neural network s. A database of 71 building projects is collected from the construction industry of the Gaza Strip. Several significant parameters were identified for the structural skeleton cost of the project and yet can be obtained from available engineering drawings and data at the pre-design stage of the project. The input layer of the Artificial Neural Networks (ANN) model comprised seven parameters, namely: ground floor area, typical floor area, number of storeys, number of columns, type of footing, number of elevators and number of rooms. The developed ANN model had one hidden layer with seven neurons. One neuron representing the early cost estimate of buildings formed the output layer of the ANN model. The results obtained from the trained models indicated that neural network s reasonably succeeded in predicting the early stage cost estimation of buildings using basic information of the projects and without the need for a more detailed design. The performed sensitivity analysis showed that the ground floor area, number of storeys, type of foundation and number of elevators in the buildings are the most effective parameters influencing the early estimates of building cost.
