Eken, G (2022) Using natural language processing for automated construction contract review during risk assessment at the bidding stage. PhD thesis, Middle East Technical University, Turkey.
Abstract
Construction sector contains various risks, and construction projects are open to failure due to their nature and the involvement of multiple parties. Contracts are legal documents that are used to define the responsibilities of parties in a project. Risks that are taken by a party are highly related to their positions in the contracts. However, contracts are prepared by clients, and contractors generally do not have enough time to review their responsibilities before preparing their bids. Moreover, contracts may not always be clear in terms of all subjects. As a result, it may lead to ambiguities. Advances in information technology (IT) may provide solutions to the construction sector in this area. Natural Language Processing (NLP) focuses on using computers to understand, process, and manipulate natural language text to achieve a variety of objectives. NLP can be used to review contract documents within seconds depending on the volume of the documents and available processing power. In this study, FIDIC standard forms of contracts were selected and all sentences were labeled with sentence type and ownership in order to create a training dataset for machine learning applications. In addition to the training dataset, the test dataset was created by using a contract of the real construction project. By using created datasets, 5 different machine learning algorithms were trained with different NLP techniques. The results of 12 machine learning models were evaluated, and the selected ones were combined by the ensemble method. In conclusion, sentence types in a FIDIC contract were categorized 89% accurately as heading, definition, obligation, risk, and right. Additionally, related parties for sentences that imply risk right and obligation were predicted 83% accurately. The proposed method can be used by contractors to quickly classify the contract text in order to identify the contractual risks required to decide risk premiums in the tender preparation phase.
| Item Type: | Thesis (Doctoral) |
|---|---|
| Thesis advisor: | Birgönül, M T and Toker, İ D |
| Uncontrolled Keywords: | FIDIC; bidding; failure; information technology; learning; machine learning; ownership; risk assessment; training |
| Index terms: | automated construction, FIDIC, standard form of contract, machine learning application, risk assessment, ownership, risk premium, machine learning, dataset, construction sector, information technology, bidding, construction project, documents, contract document |
| Subjects: | computing systems, artificial intelligence, economics, contractual condition, professional development, data management, industry analysis, standard forms of contract, contractual arrangements, bidding, economic analysis, production management, automation and robotics, financial risk |
| Topics: | Procurement, Project Management, Organizational Design, Contract Administration, Digital Applications, Business Strategy, Cost Management, Research Practice, Information Management |
| Descriptive scope: | 3 PCT |
N.B. Descriptive scope is a count of how many of the five facets of empirical research are indicated by the words used in title, abstract and keywords. It is not intended as a judgement on the research; merely a count of the kind of word we would expect to indicate Phenomenon, Concepts, Theoretical framing, Empirical techniques, Analytical techniques. If all five are present, then a code of “5 PCTEA” will indicate this. If you feel the coding for this record is questionable, we welcome discussion around the terms we matched or the way we categorized them. The facet you would expect may not be coded, or a facet may be coded inappropriately. This can also bear on a larger question, of which facets should be treated as defining in construction management research. Please get in touch, and we will look at it. More details here