A Neural Network Mark-up Estimation Model for Syrian Contractors

One of the most important decisions that have to be made by construction contractors is how much to mark-up the estimated cost of a new project. The main objectives of this paper are to model the relationship between mark-up estimation and the key factors affecting it and to compare the application of regression analysis and neural network techniques on the mark-up decision making process in order to find which technique is more reliable in terms of accuracy and robustness. The most influential mark-up factors were identified through a formal questionnaire survey conducted among Syrian contractors. Subsequently, data on one hundred and eleven real-life bidding situations was collected from Syria. Ninety-six of these projects were used to develop linear, non-linear regression and neural network mark-up models. The remaining fifteen projects were randomly held-back for validating the developed models. The neural network model proved to be robust and more accurate than the regression models. Although this study was carried out in the context of the Syrian construction industry, the methodology and the findings have much broader geographical applicability.