Application of decision tree technique to analyze construction project data

Abstract

Data mining is the analysis of data from data-warehouse using series of mathematical and statistical methods to detect patterns. Such analysis derives important information which has proved the basis of accurate decision making in retail, banks, fraud detection, customer analysis etc. Construction Industry has benefited a little from these analytical tools mainly due to unavailability of work showing the application of these techniques in analysis of data in construction. A methodology to apply decision tree to analyze the construction labor productivity is presented in the paper. The methodology addresses three areas in decision tree construction process. These include selecting more influential attributes, combining multiple attributes and defining the threshold to ignore irrelevant attributes. The methodology gives more realistic results than the traditional method of decision tree. © 2010 Springer-Verlag Berlin Heidelberg.