A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
An Empirical Analysis of Risk Similarity among Major Transportation Projects Using Natural Language Processing
Risk management is widely recognized as a best practice for public agencies to ensure the successful implementation of major transportation projects. The conventional approach to identify and evaluate project risks is dominated by getting input from subject matter experts at risk workshops. However, the uniqueness of such a risk assessment approach remains unexamined. How different are the risks among various projects? Does the risk register reflect the unique feature of a project? The goal of this study is to measure the similarity of project risks across various groups by evaluating 70 major transportation projects delivered under various methods. The similarity index is calculated at three levels, that is, the entire document of the risk register, individual risk item, and the probability and consequence of each risk using a systematic comparative analysis based on natural language processing (NLP) and a state-of-the-art deep learning algorithm named Word2vec. Our study reports a high similarity of risk registers among different projects at all three levels. The analysis does show a lower similarity of risk registers for public–private partnerships (P3) projects. The primary contributions of this study are (1) develop a new approach to analyze the risk registers at the project level as the main output of risk management practice, and (2) establish the relation of risk uniqueness and project delivery method in transportation projects. Results suggest that a data-driven approach may be possible to help project teams develop a common risk register while allowing the teams to focus on each project’s unique risks.
An Empirical Analysis of Risk Similarity among Major Transportation Projects Using Natural Language Processing
Risk management is widely recognized as a best practice for public agencies to ensure the successful implementation of major transportation projects. The conventional approach to identify and evaluate project risks is dominated by getting input from subject matter experts at risk workshops. However, the uniqueness of such a risk assessment approach remains unexamined. How different are the risks among various projects? Does the risk register reflect the unique feature of a project? The goal of this study is to measure the similarity of project risks across various groups by evaluating 70 major transportation projects delivered under various methods. The similarity index is calculated at three levels, that is, the entire document of the risk register, individual risk item, and the probability and consequence of each risk using a systematic comparative analysis based on natural language processing (NLP) and a state-of-the-art deep learning algorithm named Word2vec. Our study reports a high similarity of risk registers among different projects at all three levels. The analysis does show a lower similarity of risk registers for public–private partnerships (P3) projects. The primary contributions of this study are (1) develop a new approach to analyze the risk registers at the project level as the main output of risk management practice, and (2) establish the relation of risk uniqueness and project delivery method in transportation projects. Results suggest that a data-driven approach may be possible to help project teams develop a common risk register while allowing the teams to focus on each project’s unique risks.
An Empirical Analysis of Risk Similarity among Major Transportation Projects Using Natural Language Processing
Erfani, Abdolmajid (author) / Cui, Qingbin (author) / Cavanaugh, Ian (author)
2021-10-12
Article (Journal)
Electronic Resource
Unknown
Perceived risk of lock-in in the front-end phase of major transportation projects
| Springer Verlag | 2022
Risk analysis and management for major construction projects
| British Library Conference Proceedings | 1998