ENHANCING CONCEPTUAL COST ESTIMATION IN THE CONSTRUCTION INDUSTRY: A BAGGING-BASED ENSEMBLE APPROACH

Date

2023-9-07

Author

Razi, Taimoor

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

132
views

216
downloads

Cite This

Accurate conceptual cost estimates are significant in assessing the feasibility and budgets of construction projects. Achieving a good accuracy range in these estimations is still difficult, though, because of the scant information provided at the initial stages of construction. For cost estimation in construction, a variety of methodologies and procedures have been used; however, they often suffer from inherent limitations. In recent years, ensemble methods in machine learning have emerged as a promising approach for addressing this issue. This thesis proposes an ensemble learning approach with bagging, leveraging Support Vector Regression (SVR) and Multi-Layer Perceptron Neural Networks (MLP NN) models, to estimate conceptual costs in construction projects. By combining multiple models trained on bootstrap samples, the main objective of the proposed framework is to enhance the accuracy of conceptual cost estimates. Realworld construction datasets are used in the evaluation of the suggested method, and the evaluation metric used is the Mean Absolute Percentage Error (MAPE). Comparative analysis between the ensemble models and models without ensemble learning demonstrates the superior performance of the ensemble technique in reducing prediction errors. The findings highlight the potential of bagging in enhancing cost estimation accuracy, specifically for SVR and MLP NN models. However, limitations include the need for further research on larger datasets and alternative ensemble methods. Overall, this research contributes to advancing cost estimation in construction, providing insights for future studies and guiding practitioners in enhancing cost estimation models.

Subject Keywords

Construction Cost Estimations, Bagging, Neural Networks, Support Vector Regression

URI

https://hdl.handle.net/11511/105376

Collections

Graduate School of Natural and Applied Sciences, Thesis