The application of regression analysis and neural networks to predict elemental building costs

David Lowe, Margaret Emsley

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Ongoing research at The University of Manchester has resulted in the production of ProCost – an early stage building cost estimation tool that has the ability to produce single figure estimates following the description of the proposed building by the user. Recent research has indicated that single figure estimating is not of great use to the estimator and that an elemental cost estimating tool is required within the software’s interface. This realisation initiated the investigation into separating the ProCost output into elements. In order to do that a large database of 360 buildings with known characteristics and elemental costs was compiled. Regression analysis and artificial neural networks techniques were used to model the relationship between the building descriptive variables and their elemental costs. This paper examines the results of the regression analysis and proposes the application of neural networks due to their ability to model non-linear relationships.
Original languageEnglish
Title of host publicationProceedings of the First International Symposium on Commercial Management
EditorsDavid Lowe, Margaret Emsley
Place of PublicationUnited Kingdom
PublisherUniversity of Manchester Institute of Science and Technology (UMIST)
Pages186-194
Number of pages9
ISBN (Print)9547918-1-1
Publication statusPublished - 1 Apr 2005
Event1st International Symposium on Commercial Management 2005 - Manchester, United Kingdom
Duration: 7 Apr 20057 Apr 2005

Conference

Conference1st International Symposium on Commercial Management 2005
Country/TerritoryUnited Kingdom
CityManchester
Period7/04/057/04/05

Fingerprint

Dive into the research topics of 'The application of regression analysis and neural networks to predict elemental building costs'. Together they form a unique fingerprint.

Cite this