摘要
目的/意义探索并实现儿科医学试题答案的自动化解析,提高试题答案解析编撰效率与质量。方法/过程提出一种隐性语义索引、MC-BERT和CoSENT模型相结合的方法。首先使用基于隐性语义索引的方法和MC-BERT模型从参考文档中抽取多个候选答案解析,然后利用CoSENT模型计算候选解析、试题题干和答案选项之间的相似度,选取相似度最高的候选解析作为最终答案解析。结果/结论该方法答案解析精确率达到72.6%,相较单一方法或模型明显提高查全率和精确率,有效提高了编撰试题答案解析的效率,减轻教育工作者负担,并可为教育研究提供重要的数据支持。
Purpose/Significance To explore and implement the automated interpretation of pediatric medical exam answers,so as to enhance the efficiency and quality of answer explanation compilation.Method/Process The paper proposes a method that combines latent semantic indexing(LSI),MC-BERT,and the CoSENT model.Initially,multiple candidate answer explanations are extracted from reference documents using the LSI method and the MC-BERT model.Subsequently,the CoSENT model is employed to calculate the similarity between the candidate explanations and the question stems as well as the answer options.The candidate explanation with the highest similarity is then selected as the final answer explanation.Result/Conclusion The experimental results show that the method presented in this paper achieves a precision rate of 72.6%.Compared to single methods or models,it significantly improves the recall and precision of answer parsing,effectively enhances the efficiency of compiling question answer explanations,reduces the burden on educators,and provides significant data support for educational research.
作者
王娟
侯丽
孙月萍
李佳明
杨丽
董良广
李云汉
WANG Juan;HOU Li;SUN Yueping;LI Jiaming;YANG Li;DONG Liangguang;LI Yunhan(Institute of Medical Information,Chinese Academy of Medical Sciences&Peking Union Medical College,Beijing 100020,China;People’s Medical Publishing House Co.Ltd.,Beijing 100021,China;Beijing Institute of Surveying and Mapping,Beijing 100038,China)
出处
《医学信息学杂志》
CAS
2024年第10期11-17,共7页
Journal of Medical Informatics
基金
中国医学科学院医学与健康科技创新工程项目(项目编号:2021-I2M-1-001)
医学融合出版知识技术重点实验室。
