Enhancing the accuracy of construction project completion estimates by incorporating updated timing-based error margins (utems)

The Earned Value Method (EVM) has been used to predict project performance for many years, with numerous studies confirming its reliability and highlighting its advantages and limitations. However, using simple methods to predict highly reliable project outcomes remains crucial for engineers and managers during project implementation. This study aims to use regression methods to model the impact of update timing on error margins, referred to as UTEMs, for estimates at project completion. Data were collected from five residential projects of varying sizes, processed, and standardized before statistical analysis. The regression models include linear, exponential, polynomial, logarithmic, and power functions. The research results show that the exponential regression model has an excellent fit, with an R-squared value of 95.6%. The accuracy of the proposed model is high, with MPE = 3.3%, MSE = 0.3%, RMSE = 5.2%, and MAPE = 13.7%. The sensitivity analysis results indicate maximum and minimum error margins of 76.5% and 2.6% at the project's start and end, respectively. The proposed model's prediction results were verified using an actual project, demonstrating minimal deviations. The findings of this study are anticipated to add value to the broader understanding of EVM extensions and assist managers in accurately predicting project outcomes.