Universiti Teknologi Malaysia Institutional Repository

Predicting the performance of traditional general contract projects: A neural network based approach

Mohamad Zin, Rosli and Mansur, S. A. and Bakri, A. and Tan, Caren Cai Loon (2006) Predicting the performance of traditional general contract projects: A neural network based approach. In: 6th Asia-Pacific Structural Engineering and Construction Conference, 5-6 September 2006, Kuala Lumpur, Malaysia.


Official URL: http://civil.utm.my/apsec2015/


Several studies had shown that many project managers are facing difficulties in predicting the performance of Traditional General Contract (TGC) projects. This is due to the fact that there are many factors that affect TGC project success. This paper presents the TGS project success factors that have been identified. In addition, a model to predict the performance of TGC project based on time is also described. Through literature research, a total of forty-four factors affecting TGC project success had been established. The degree of importance for these factors was determined through questionnaire survey. The outcome of the survey formed a basis for the development of the project performance prediction model. The best model was found to be a multi-layer back-propagation neural network consists of eight input nodes, five hidden nodes and three output nodes. The model was tested by using data from nine new projects. The results showed that the mean error for this prediction model is relatively low. The model enables all parties involved in TGC projects to predict and ensure that their project performance is within the time constraints.

Item Type:Conference or Workshop Item (Paper)
Uncontrolled Keywords:artificial neural network, project performance, traditional general contract
Subjects:T Technology > TA Engineering (General). Civil engineering (General)
Divisions:Civil Engineering
ID Code:549
Deposited By: Norazila Safri
Deposited On:16 Feb 2007 07:59
Last Modified:23 Aug 2017 00:31

Repository Staff Only: item control page