Budget and cost contingency CART models for power plant projects

Md Arifuzzaman, Uneb Gazder, Muhammad Saiful Islam*, Martin Skitmore

*Corresponding author for this work

Research output: Contribution to journalArticleResearchpeer-review

1 Citation (Scopus)
128 Downloads (Pure)


Cost overruns are a ubiquitous feature of construction projects, and realistic budgeting at the development stage plays a significant role in their control. However, the application of existing models to budgeting for power plant projects is restricted by the limited amount of project-specific cost data available. This study overcomes this by using a Classification and Regression Tree (CART) approach involving mixed methods of website visits, document study, and expert opinion to predict the amount of project cost (PC) and cost contingency (CC) needed to cover probable cost increases by the use of models containing readily available project attributes and national economic parameters at the project development stage. The modeling process is demonstrated and tested with a case study involving 58 Bangladeshi power plant projects – producing average absolute errors ranging from 0.7% to 1.7% and enabling project cost, inflation rate, and GDP to be identified as significant factors affecting PC and CC modeling. The approach can be applied to predict the PC during preliminary budgeting and selecting a project type and location aligned to the country’s economic status and policy-making strategies, thus facilitating further investment decisions.

Original languageEnglish
Pages (from-to)680-695
Number of pages16
JournalJournal of Civil Engineering and Management
Issue number8
Publication statusPublished - 21 Sept 2022


Dive into the research topics of 'Budget and cost contingency CART models for power plant projects'. Together they form a unique fingerprint.

Cite this