Construction industry is overwhelmed by increasing number and severity of disputes. The primary objective of this research is to predict the occurrence of disputes by utilizing machine learning (ML) techniques on empirical data. For this reason, variables affecting dispute occurrence were identified from the literature and a conceptual model was developed to depict the common factors. Based on the conceptual model, a questionnaire was designed to collect empirical data from experts. Chi-square tests were conducted to reveal the associations between input variables and dispute occurrence. Alternative classification techniques were tested, and support vector machines (SVM) classifier achieved the best average accuracy (90.46%). Ensemble classifiers combining the tested classification techniques were developed for enhanced prediction performance. Experimental results showed that the best ensemble classifier, obtained from majority voting technique, can achieve 91.11% average accuracy.Based on Chi-square tests, the most influential factors on dispute occurrence were found as variations and unexpected events in projects. Other important predictors were all related to skills of the parties involved. This study contributes to the construction dispute domain in three ways (1) by proposing a conceptual model that combined the diverse efforts in the literature for identifying variables affecting dispute occurrence, (2) by highlighting the influential factors such as response rate and communication skills as indicators for potential disputes, (3) by providing an empirical ML-based model with enhanced prediction capabilities that can function as an early-warning mechanism for decision-makers.
This paper compares classification performances of machine learning (ML) techniques for forecasting dispute resolutions in construction projects, thereby mitigating the impacts of potential disputes. Findings revealed that resolution cost and duration, contractor type, dispute source, and occurrence of changes were the most influential factors on dispute resolution method (DRM) preferences. The promising accuracy of the majority voting classifier (89.44%) indicates that the proposed model can provide decision-support in identification of potential resolutions. Decision-makers can avoid unsatisfactory processes using these forecasts. This paper demonstrated the effectiveness of ML techniques in classification of DRMs, and the proposed prediction model outperformed previous studies.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.