Understanding The Multiple Functions Of International Construction Contracts: Based On Natural Language Processing

International construction contracts are complex and cumbersome.Whether it is contract research or contract management,it is necessary to establish a comprehensive understanding of complex contracts.This research focus on a multi-functional perspective to classify the text of international construction contract conditions,in order to gain a more comprehensive analysis of international construction contracts.The particular conditions are sensitive to the characteristics of the project,leading to extremely high cost of researching it.Therefore,contract research has rarely included particular conditions.The content analysis method for text classification is not suitable for particular conditions with high specificity.This research uses four Machine Learning algorithms to develop a contract function identification model and realizes automatic text classification.Using the contract function identification model,this research collected 9 sets of particular conditions for each function category data,and performed function complexity analysis.The analysis results show that: the functional complexity of particular conditions differs from each other,but the complexity distribution of each function category of particular conditions is similar to that of general conditions;by defining four types of sentences,particular conditions directly influence the complexity of the general condition function and then the overall function complexity of the contract conditions;the particular conditions positively affect the overall function complexity of the contract conditions,among which the influence on the coordination function is the greatest,the control is the second,and the adaptation is the least.