科学探究学习中思维导图的自动评价方法

陈娟娟¹ 郭 阳¹ 张子柯²

（1.浙江大学 教育学院，浙江 杭州 310058；2. 浙江大学 传媒与国际文化学院，浙江 杭州 310058）

【摘 要】科学探究学习提倡学生在实践中通过观察、推理构建知识，深入理解科学现象的规律与本质。将思维导图应用于科学探究学习中可以将学生的思维可视化，既是有效的学习脚手架，又是重要的系统性思维评价工具。但目前研究中对思维导图的评价仍高度依赖人工，存在着主观性强、评分效率低等问题。为提高思维导图的评价效率与准确率，本研究探讨了现有的科学探究学习中思维导图的自动评价方法与其局限性。现有的思维导图自动评价方法分为三类：拓扑评分法、专家参考图比较法和综合分析法。其中拓扑评分法无法识别思维导图中命题的准确性；专家参考图比较法为保证其准确性会要求学生尽量使用与专家相同的词汇表达命题，限制了学生的思考；综合分析法在两者之上进行了改进但也不能完全取代人工评价。思维导图自动评价算法的后续研究首先应建立统一的评价指标体系，综合不同算法的优势构建在指标体系下各个评价维度都更为优秀的评价模型。其次，需要进一步细化评价结果，提高评价的精度和全面性，为学生在科学探究的过程中提供精细化的反思性反馈。最后，未来的研究可聚焦采用最先进的人工智能算法搭建模型，这将有助于思维导图自动评价算法的进一步发展和应用。

【关键词】思维导图；自动评价；科学探究学习；科学教育

A Review of Automatic Evaluation Methods of Mind Mapping in Scientific Inquiry Learning

CHEN Juanjuan¹, GUO Yang¹ and ZHANG Zike²

（1. College of Education, Zhejiang University, Hangzhou 310058, China ; 2. College of Media and International Culture, Zhejiang University, Hangzhou 310058, China）

Abstract: Scientific inquiry learning emphasizes student knowledge building through observation and reasoning to deeply understand scientific phenomena. Mind mapping acts as an effective learning scaffold and an important evaluation tool in visualizing students’thinking However, studies pay insufficient attention to the evaluation of student-constructed mind mapping, and the evaluation of mind mapping is still highly dependent on manual work, resulting in strong subjectivity and low scoring efficiency. Currently three types of methods for automatic evaluation of mind maps are uesd, including topological scoring, expert reference map comparison and comprehensive comparison methods. Topological scoring cannot identify the accuracy of propositions in mind maps, and expert reference map comparison method requires students to use the same vocabulary as experts to express propositions to ensure its accuracy, which limits students’thinking. The comprehensive scoring method is an improvement, but it cannot completely replace the manual evaluation. The future research of automatic evaluation should first establish a unified automatic evaluation index system and synthesize different algorithms to build a more excellent evaluation model for each index. It is necessary to further refine the evaluation results, to improve the accuracy of the evaluation, and to provide refined reflective feedback for students. Finally, future research could focus on building scoring models using the most advanced AI algorithms.

Keywords: mind mapping; automatic scoring; scientific inquiry learning; science education

下载： 科学探究学习中思维导图的自动评价方法.pdf