Development of Automatic-Extraction Model of Poisonous Clauses in International Construction Contracts Using Rule-Based NLP

Abstract

As construction projects have significantly increased in size and become more complicated, the number of claims and dispute cases between participating parties during the construction work have been continuously increasing. To prevent such claims and disputes, the participants need to be assured of their contractual positions and rights based on contract facts. For this reason, the process of writing and reviewing the contracts for construction work is crucial. Most international construction projects require contract management teams to review all the possible risks in the contracts during the bidding periods. However, it is very difficult to review a vast number of contracts in a short period of time. Therefore, in this study, we proposed an automatic model of contract-risk extraction based on natural language processing (NLP) that can automatically detect the poisonous clauses of the contract in order to support contract management for construction companies (contractors). In validating the performance of the automatic model developed in this study, we found that the precision and recall were both 81.8% compared with manual review. This study is meaningful since a model has been developed that can carry out a preemptive contract-risk review.