Factor selection for delay analysis using Knowledge Discovery in Databases

Today's construction project has become a very complex, high-risk, multiparty endeavor. Construction projects are composed of many interrelated elements of labor, cost, material, schedule, and other resources, making it difficult to discern which factors were the main causes for delay on a given project. Were all relevant factors to be considered, it would become an overwhelming task. On the other hand, it would be very difficult to know on which factors to focus if only a limited number of factors were to be considered. This paper presents a methodology for factor selection; identifying which factors in an on-going construction project contribute most to the experienced delays. Factor selection is defined as the process of finding relevant factors among a large set of original attributes with the objective of best representing the original dataset and utilizes Knowledge Discovery in Databases (KDD), which is a data analysis process to discover useful knowledge in a large database. A specific construction project has been analyzed to identify main factors of construction delays through the process of statistical measurements and machine learning algorithms.