基于RoBERTa的油气高风险作业管控措施分类方法

Classification Method for Control Measures of Oil and Gas High-risk Operations Based on RoBERTa

本文提出了一种基于RoBERTa的油气高风险作业管控措施分类方法,旨在解决高风险作业JSA库针对性不足的问题,该方法可以自动、准确地将管控措施划分成关键性措施与一般性措施,能够帮助使用者快速找到并采取关键的风险管控措施,提升JSA库的应用效果和风险管控水平。此外,本文还对管控措施分类结果进行了实验验证,结果表明该方法具有较高的准确性和可靠性,F1值达到76%,能够满足实际应用的需要。

This paper proposes a classification method for controlling measures of oil and gas high-risk operations based on RoBERTa,aiming to solve the problem of insufficient pertinence of risk control measures in high-risk operations JSA library.The method can automatically and accurately classify control measures into key measures and general measures,helping users quickly find and take key risk control measures,and improving the application effect and risk control level of the JSA library.In addition,this paper also verifies the classification results of control measures through experiments,and the results demonstrate that this method has high accuracy and reliability,with an FI value of 76%,which can meet practical applications.