Construction knowledge discovery system using fuzzy approach

Most research works in simulating construction operations have predominantly focused on modeling and mistreated data preparation that is paramount for simulation. To prepare data for simulation process, a knowledge discovery system (KDS) is indispensable in extracting hidden knowledge from construction data sets. This knowledge is typically hard to obtain using traditional means, such as statistical analysis. The presented research develops, using fuzzy approach, a KDS to prepare, utilize, analyze, and extract the hidden patterns from construction data to predict work task durations. The KDS depends mainly on finding the relation between quantitative and qualitative variables, which affect the duration of construction operations and work tasks as well as prepare data for simulation modeling. It consists of two stages: data processing and mining. Data processing consists of cleaning, integrating, transforming, and selecting the appropriate knowledge. Data mining consists of selecting the factors that affect a construction operation, generating their fuzzy sets, defining fuzzy rule and models, developing a fuzzy knowledge base, and testing the effectiveness of this knowledge base in predicting work task durations. The developed KDS has been tested using a construction case study in which the results found satisfactory with an average validity percent of 92%. The developed system assists researchers and practitioners in utilizing historical construction data to extract knowledge that could not be obtained by traditional techniques and precisely predicting work task durations.