Abstract
Ongoing research at UMIST commenced in 1997 has resulted in the production of ProCost, an early stage building cost modelling tool. The software is based on Artificial Neural Network technology to produce single figure estimates of the total building cost. Recent research has however indicated that cost estimators cannot nowadays rely on single figure cost estimating techniques (Soutos & Lowe, 2003 and Soutos & Lowe, 2004). This initiated the next stage of the research, which involved the investigation of subdividing the single figure cost output into a cost for each building element. In order to proceed, a large database of 360 buildings with developed elemental cost breakdowns was formulated with the aid of Building Cost Information Service (BCIS). This database was used to investigate the way that a series of building characteristics affect the cost of building elements. In order to model these relationships, linear regression analysis was used. The results of this method are discussed in this paper. Artificial Neural Networks, are then proposed as an additional way of modelling the data, and their advantages over regression analysis are considered. This paper presents the results of an extensive piece of research with respect to ProCost and discusses its evolution into a powerful cost estimating package.
Original language | English |
---|---|
Title of host publication | Computing in Civil Engineering 2005 |
Editors | L. Soibelman, F. Pena-Mora |
Publisher | American Society of Civil Engineers |
Pages | 1503-1514 |
Number of pages | 12 |
ISBN (Print) | 0784407940 |
DOIs | |
Publication status | Published - 2005 |
Event | 2005 ASCE International Conference on Computing in Civil Engineering - Cancun, Mexico Duration: 12 Jul 2005 → 15 Jul 2005 |
Conference
Conference | 2005 ASCE International Conference on Computing in Civil Engineering |
---|---|
Country/Territory | Mexico |
City | Cancun |
Period | 12/07/05 → 15/07/05 |
Keywords
- Artificial neural networks
- Cost estimating
- Cost modelling
- Elemental estimating
- Regression analysis
ASJC Scopus subject areas
- General Engineering