Cost Estimation of Building Construction Projects in Gaza Strip Using Support Vector Machines Model (SVM)

Abstract

Construction industry is considered an important sector for the development in Gaza Strip. Estimating process is an important element of the construction industry. Early stage cost estimate plays a significant role in the success of any construction project. The research aims to predict the parametric cost estimation in construction building projects in Gaza Strip using support vector machine model. This research has been conducted through literature reviews of the cost estimating process, followed by a field survey done by structured interviews to define the factors which effect the parametric cost estimation of building construction projects. The study shows that nine factors are affecting construction cost. The main factors are the area of typical floor, numbers of stories and the building type. Eighty case studies from real executed construction projects in Gaza Strip were collected to build up SVM model. Neurosolution software version 6.07 was used to train the models. The results of the trained models indicated that SVM reasonably succeeded in estimating the cost of building projects without the need for more detailed drawings. The average error of test dataset for the adapted model was 2.3% which is largely acceptable. One of the main recommendations of this research is to join the developed model with cost index to give an accurate estimate in any time. In addition, it encourages all parties involved in construction industry to pay more attention for developing model in cost estimation by archiving all projects data, and conducting more studies and workshops to obtain maximum advantage of this new approach.