摘要：
目的 为提高撤稿记录在科技伦理风险识别与案例学习中的利用价值，构建科技伦理参考学习库，并从撤稿案例解析可靠性和用户接受度分别开展效能评估。 方法 基于本体论思维，结合政策规范梳理与撤稿声明分析，建立6类撤稿原因、20项违规特征分类体系；采集Crossref撤稿记录，利用DeepSeek进行语义解析与分类，整合政策规范等资源形成资源中心；通过人工标注验证撤稿案例解析可靠性，并基于技术接受模型对100名科技期刊编辑开展问卷评价。 结果 参考学习库结构化归类2403条科技伦理相关撤稿案例，撤稿原因和违规特征识别F1值分别为93.86%和90.41%，问卷各维度核心题项的积极反馈比例均达86%及以上。 结论 参考学习库可为科技期刊开展科技伦理风险识别、风险处置和风险防范学习提供支持。
关键词:
科技伦理,
撤稿案例,
参考学习库,
本体论思维,
技术接受模型,
效能评估
Abstract:
Purposes To enhance the value of retraction records in science and technology ethics risk identification and case-based learning， this study constructs a science and technology ethics reference learning repository and evaluates its effectiveness from the perspectives of the reliability of retraction case parsing and user acceptance. Methods Based on ontological thinking， combined with policy and regulatory review and analysis of retraction statements， a classification framework comprising 6 categories of retraction reasons and 20 misconduct features was established. Retraction records were collected from Crossref， and DeepSeek was used for semantic parsing and classification. Policy， regulatory， and other related resources were integrated to form a resource center. Manual annotation was used to verify the reliability of model parsing， and a questionnaire survey based on the Technology Acceptance Model was conducted among 100 editors of scientific and technical journals. Findings The reference learning repository structured and classified 2403 science and technology ethics-related retraction cases. The F1 values for identifying retraction reasons and misconduct features were 93.86% and 90.41%， respectively， and the positive feedback rates for the core items across all questionnaire dimensions reached 86% or higher. Conclusions The reference learning repository can support scientific and technical journals in science and technology ethics risk identification， risk management， and risk prevention learning.
Key words:
Science and technology ethics,
Retraction case,
Reference learning repository,
Ontological thinking,
Technology acceptance model,
Effectiveness evaluation
苗婧, 周国栋, 贾丽丽, 杨鹏, 邹文婕, 张昕, 尹欢. 基于撤稿案例的科技伦理参考学习库建设与效能评估[J]. 中国科技期刊研究, 2026, 37(5): 611-622.
MIAO Jing, ZHOU Guodong, JIA Lili, YANG Peng, ZOU Wenjie, ZHANG Xin, YIN Huan. Construction and effectiveness evaluation of a reference learning repository for science and technology ethics based on retraction cases[J]. Chinese Journal of Scientific and Technical Periodicals, 2026, 37(5): 611-622.