【作 者】：
张义兵，高兴启，满其峰，胡金艳，许馨予
【关 键 词 】：
知识建构； 教育知识图谱； LDA； 跨社区
【栏 目】：
学习环境与资源
【中文摘要】：
跨越社区边界的知识共享是计算机支持下的协作学习的知识创造来源之一，这对重视知识创造过程的知识建构教学来说尤其重要。如何建立跨越边界的知识建构元空间，成为近年来国际跨知识建构社区信息交互研究的重要问题之一；但由于其实现过程受到人工规划、设计与实现的约束，多处于搭建概念模型的研究阶段。本研究基于教育知识图谱的理念，运用LDA主题抽取算法抽取知识建构社区的核心主题、发现跨社区的观点之间的联系，进而采用可视化的知识图谱形式构建知识建构元空间，并运用计算机实验加以验证。实验结果表明，教育知识图谱形式下的知识建构元空间在跨社区的观点的核心主题抽取方面以及跨社区观点之间的关系与呈现的功能上表现优秀，主题抽取的准确率为94.29%，观点与主题之间的关系抽取准确率为92.85%，同一主题下的观点关系的平均准确率为89.79%。知识建构元空间的建构尝试，不仅可以为国际知识建构相关研究探索新路径，也可以在实践上更好地支持跨社区的信息交互与共享。
【英文摘要】：
Knowledge sharing across community boundaries is one of the sources of knowledge creation in computer-supported collaborative learning, which is especially important for knowledge construction teaching and learning that values the process of knowledge creation. How to establish a knowledge building meta-space across boundaries has become one of the important issues in international research on information interaction across knowledge building communities in recent years. However, due to the constraints of manual planning, design and implementation, it is mostly at the research stage of building conceptual models. Based on the concept of educational knowledge graph, this study uses the LDA topic extraction algorithm to extract the core topics of knowledge construction communities, finds the connections among the views across communities, and then constructs a knowledge construction meta-space in the form of a visual knowledge map, and validates it through computer experiments. The experimental results show that the knowledge construction meta-space in the form of educational knowledge graph performs well in extracting the core topics of cross-community views and in the function of the relationship and presentation between cross-community views. The accuracy of topic extraction is 94.29%, and the accuracy of the relationship between views and topics is 92.85%. The average accuracy of viewpoint relationships under the same topic is 89.79%. The attempt to construct knowledge construction meta-space can not only explore new paths for international knowledge construction-related research, but also better support cross-community information interaction and sharing in practice.