首页|基于情绪分析的生产安全事故政府责任公众感知偏差研究

基于情绪分析的生产安全事故政府责任公众感知偏差研究

扫码查看
原文链接
  • NETL
  • NSTL
  • 万方数据
为强化政府安全生产监督及行政问责公正机制,提升生产安全事故协同治理能力,利用文本挖掘技术从个体和场域2 个层面出发,探索生产安全事故政府责任的公众感知偏差形成机理和影响因素.通过公众责任感知双向编码转换器(BERT-PPR)预测事故微博评论的情绪和归责类型,对比事故调查结果得到政府责任的公众感知偏差,并基于二元逻辑回归考察事故信息和微博报道对感知偏差的影响.研究结果表明:采用政府形象框架以及调查结果公布阶段引发政府舆情危机的风险更高;责任人宣判阶段公众更易误判政府有责.行业、阶段、等级、形式、框架因素对生产安全事故政府责任公众感知偏差的影响不同,应采取对应措施,进而纠正相关偏差.研究结果可为安全生产领域内相关政策调整提供参考.
Research on public perception bias of government responsibility for production safety accidents based on emotion analysis
To enhance the governmental mechanism for work safety supervision and fair administrative accountability,and im-prove the collaborative governance capability of production safety accidents,the text mining technology was applied to explore the mechanism and influencing factors of public perception bias of government responsibility for production safety accidents from individual and domain levels.The bidirectional encoder representations from transformers for public perception of respon-sibility(BERT-PPR)model was utilized to predict the emotion and attribution types in microblog comments related to acci-dents.By comparing with the accident investigation results,the public perception bias of government responsibility was ob-tained,and the binary logistic regression was employed to examine the impact of accidents information and microblog reports on the perception bias.The results show that there is higher risk of triggering the government public opinion crisis by adopting the government image framework and the publication stage of investigation results.During the responsible person's judgment stage,the public is more inclined to mistakenly believe that the government is responsible.The factors such as industry,stage,grade,form and framework have different influence on the public perception bias of government responsibility for work safety accidents,and the corresponding measures should be taken to correct the relevant bias.The research results can provide refer-ence for the adjustment of relevant policies in the field of production safety.

production safety accidentgovernment responsibilityperception biaspublic opinion governanceemotion analysisbidirectional encoder representations from transformers(BERT)

张羽、周旭、梁琦

展开 >

西南石油大学 土木工程与测绘学院,四川 成都 610500

生产安全事故 政府责任 感知偏差 舆情治理 情绪分析 双向编码转换器(BERT)

四川省社科规划项目四川省科技厅软科学研究项目

SC21B1622023JDR0299

2024

10.11731/j.issn.1673-193x.2024.08.027

中国安全生产科学技术
中国安全生产科学研究院

中国安全生产科学技术

CSTPCD北大核心
影响因子：1.119
ISSN：1673-193X
年,卷(期)：2024.20(8)
  • 13