Litigation prediction model for construction disputes caused by change orders

It is all to clear that using litigation as the ultimate dispute resolution during construction projects is counterproductive. Project change is a very significant factor triggering potential enormous disputes that requires settlement through legal process. Change is still however inevitable, and construction litigation ensues on most construction projects. To predict the potential probability of construction litigation instigated by substantial change disputes, this research had three main objectives: (1) Study and investigate the impact of project change at the macro level. (2) Compare non-litigated with litigated construction project data to determine specific factors impacted by project change. (3) Develop a model to predict the probability of litigation by change disputes.;Commencing with a comprehensive literature review and an investigation of historical construction litigation filed in the Supreme Court of Wisconsin was conducted. Impact factors and the characteristics in projects causing change disputes were identified. Subsequently, an extensive data collection was completed following a set criteria for investigation. Through the testing of statistical hypothesis and feature reduction, the features (factors) utilized to develop the predictive model were determined. Two features are deemed significant: project size and percent change.;The Litigation Prediction Model (LPM) was developed utilizing the concept of the pattern classification integrated with two distinct algorisms: K-mean nearest neighbor and back-propagation neural network. A high classification rate, 90%, indicated that the model performs in an acceptable range for predicting the trend of litigation produced by change disputes.;An evaluation for the LPM was completed with an in-depth analysis for model results. This research concludes that LPM is reliable in predicting litigation in construction projects based on project size and change. With data trimming techniques applied to the LPM database, LPM can produce more accurate litigation predictions for construction projects. Finally, based on model feasibility, the LPM is designed for flexibility, to meet the demands for diverse construction projects.