Using Data Mining Algorithms to Identify Critical Patterns among Various Performance Indicators of Bundled Projects Delivered under the Construction Manager/General Contractor (CM/GC) Project Delivery Method

Project bundling is an innovative project delivery approach that groups several infrastructure projects into one contract, generally for rehabilitating or replacing them. Project bundling has numerous opportunities, such as saving costs and expediting project delivery. Although previous research presented guidelines for project bundling in terms of its opportunities, challenges, and decision-making factors, little to no studies have presented guidance in relation to its performance aspects. Also, no previous effort was conducted to offer guidance in relation to using the Construction Manager/General Contractor (CM/GC) delivery method for bundled infrastructure projects. Thus, this research develops a data mining machine learning model that discovers critical associations between the applicability levels of project performance aspects of bundled projects delivered using the CM/GC method. This research first developed a survey to collect expert opinions on project bundling performance aspects under the CM/GC method. Second, the data collected from the survey was preprocessed and transformed into the appropriate format for data mining purposes. Third, data mining algorithms were implemented to discover hidden patterns between the different project performance aspects. The results identified five critical or key associations between different performance aspects of bundled projects under the CM/GC method. Ultimately, this research helps decision-makers target the most critical aspects of project bundling and thus optimize the bundling practices of their construction and infrastructure projects under alternative, innovative project delivery methods.