Abstract:
The purpose of this research is to provide an efficient and effective mean of estimating overhead costs of construction in Pakistan as a percentage of the total costs. In order to achieve this, applications of artificial neural networks were implied. This research presents an in depth knowledge over the main factors affecting the overhead costs and then using those factors to develop a neural network model which would help predict overhead cost of the building projects of Pakistan. The main data collection was done through telephonic interviews with the contractor personnel and asking them about their views on the shortlisted factors in light of overhead costs. The main data set was then used to design, test and validate the neural network model. The model that presented best results was than selected as a predicting tool for the estimation of overhead costs.
