Models for early cost estimating using linear regression: penitentiary projects modeling

Abstract Early cost estimation of construction projects is not an easy task, as such projects involve a high level of inaccuracy and uncertainty. Even in the early stages, errors in the estimates can result in financial loss and jeopardize construction completion. Therefore, the main objective of this research study is to present a framework for building cost estimation models, using the linear regression technique. The framework method is divided into five phases: (1) identifying the model’s requirements, (2) selection of the independent variables, (3) database construction, (4) data modelling and (5) model performance evaluation. A case study was conducted on federal penitentiary construction projects to test the applicability of the framework. Through the case study, two valid models were built, and their margins of error were 23 and 25%. The framework itself is one of the main contributions of this study, and it can be replicated by practitioners to develop models for construction cost estimation.