A1056
Title: Final project cost estimation in EVMS using LMM
Authors: Yonghyun Um - Chungnam National University (Korea, South) [presenting]
Min Koo Lee - Chungnam National University (Korea, South)
Abstract: The earned value management system (EVMS) is a management tool utilized by organizations to track costs and schedules periodically. This enables them to assess the current status and predict the future of their projects. While various existing methods exist for estimating final project costs, they often lack accuracy and efficiency due to limited data utilization. The aim is to propose a novel approach for estimating final project costs by leveraging data from completed projects to predict ongoing ones. Specifically, a linear model must be developed to increase the accuracy of the final project cost estimate and minimize the variance of the error. To achieve this, the dynamic time warping (DTW) algorithm is employed to cluster time-series data such as actual cost (AC) and planned value (PV). Subsequently, analysis was conducted using a linear mixed model (LMM). An LMM was constructed using a variable combining the clusters of AC and PV as a random effect. All covariates expected to affect the final project cost, which is the response variable, were added, and backward elimination was performed. Through a model created using data from completed projects, the differences is examined in estimated values from existing final project cost estimation methods.
