期刊文献+

基于RoBERTa模型的驾驶培训学员投诉数据分析

RoBERTa Model Based Driving Trainee Complaints Analysis
在线阅读 下载PDF
导出
摘要 为了从需求侧找准机动车驾驶培训行业存在的问题症结,促进机动车驾驶培训行业健康发展,基于RoBERTa模型对2022—2023年政务服务便民热线和交通运输服务监督热线记录的50余万条驾驶培训学员投诉文本数据展开分析。首先,运用正则表达式等方法进行学员投诉数据预处理。其次,采用专家经验法对学员投诉数据反映的问题及原因进行分类标注和验证,生成训练数据集,基于RoBERTa构建学员投诉数据分类模型,并训练确定模型的关键参数。再次,基于数据分类模型对学员投诉反映的问题及原因进行统计分析。最后,提出了推动培训合同标准化,加强驾校收(退)费管理和教练员队伍素质教育,以及推动数字化转型等措施建议。结果表明,驾校在培训收(退)费、培训服务、经营管理、招生、学时记录、合同使用、教练员教学服务等方面存在的问题较为突出;学员投诉问题对应的投诉率在不同地区的表现基本一致,但学员投诉问题对应的原因在不同地区存在差异。 To accurately identify the crux of problems in the motor vehicle driver training industry from the demand side,and promote its healthy development,this study analyzed over 500000 trainee complaint records from government service hotlines and transportation supervision hotlines from 2022 to 2023 by using RoBERTa model.First,the regular expressions and other methods were applied to preprocess the complaint data.Second,the expert experience method was used to classify and label the problems and causes reflected in the complaints,generating a training dataset.The complaint classification model was constructed based on RoBERTa with key parameters determined through training.Third,the problems and reasons reflected in the complaints from trainees were statistically analyzed based on the data classification model.Finally,it was suggested to promote the standardization of training contracts,strengthen the management of driving school fees or refunds and quality education of coaching staff,and promote digital transformation.The result indicates that training schools have prominent problems in the issues,e.g.,fee collection and refunds,training services,operational management,student recruitment,training hour recording,contract usage,and instructor teaching services.The reasons corresponding to the issues of trainee complaints vary in different regions.
作者 朱艳红 曾诚 孟兴凯 罗文慧 ZHU Yan-hong;ZENG Cheng;MENG Xing-kai;LUO Wen-hui(Research Institute of Highway,Ministry of Transport,Beijing 100088,China;Key Laboratory of Operation Safety Technology on Transport Vehicles,Beijing 100088,China)
出处 《公路交通科技》 CSCD 北大核心 2024年第S1期453-460,共8页 Journal of Highway and Transportation Research and Development
基金 中央级公益性科研院所基本科研业务费专项资金项目(2023-9007)
关键词 交通管理 数据分类分析方法 RoBERTa模型 机动车驾驶培训 大数据分析 traffic management data classification analysis method RoBERTa model motor vehicle driver training big data analysis
作者简介 朱艳红(2001-),女,湖南浏阳人,硕士研究生.(yh.zhu@rioh.cn);通讯作者:曾诚(1978-),男,湖南娄底人,硕士研究生,研究员.(c.zeng@rioh.cn)
  • 相关文献

参考文献12

二级参考文献60

共引文献81

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部