学习分析主题结构研究及可视化分析

Topic Structure and Visualization Analysis of Learning Analytics Research

学习分析作为教育大数据的重要应用领域,自问世以来便迅速得到众多学者的广泛关注。为进一步探析学习分析研究主题,本研究以Web of Science核心库为数据来源,以2010-2015年间发表的674篇学习分析文献为研究对象,采用多种分析工具及可视化技术,从高被引文献、核心作者分布、论文高产机构以及高频关键词方面入手,梳理并揭示了当前学习分析研究现状。研究利用聚类分析等多元统计分析方法,以高频关键词相异矩阵为基础,介绍了学习分析框架模型、驱动力、方法工具、技术支持以及应用研究等五大领域的研究趋向;通过绘制战略坐标图,明确了应用研究在该领域的关键位置,并通过主题演化分析进一步梳理学习分析的研究方向,为后续相关研究提供参考。

As an important part of educational data mining, learning analytics has drawn extensive attention among international scholars since its origin. To further explore the research on learning analytics, this paper, based on 674 articles on learning analytics between 2010 and 2015 from Web of Science, explored and revealed the international research status about learning analytics. In this study, we used the ＂Bicomb＂ word frequency analysis software, the ＂Citespace＂ citation analysis tool and ＂SCI2＂ tool and visualization technology to analyze the high-cited articles, the distribution of core authors, high prolific institutions and high frequency keywords. In addition, based on dissimilarity matrix of high frequency key- words, this paper further clarified the research scope using cluster analysis. Furthermore, by mapping out the strategic diagram graph, the trend of the development orientation of learning analytics was further clarified in order to provide reference and suggestions for international research and practices on learning analytics. Results showed that the research topics had been greatly expanded and were more abundant during the 2011-2014 period. The learning analytics on behalf of communities suddenly rose and ranked to the top for four years, which made it the hottest research area during those period of times. Learning analytics showed an increasing trend year by year, mainly related to information retrieval and education data mining. While, there is a split phenomenon of the visual analytics. It has two successors involving ＂Visual analytics＂ and ＂Machine learning＂, respectively as the repre- sentative of the theme. The two successors with the learning analytics become the three core research hotspots in 2015. From the results of the study, we can conclude that the learning analytics technology is still in the initial stage and the field of research is not balanced. Learning analytics research also face many challenges concerning data security, ethics, and privacy. In general, although learning analytics technology has great application value and development potential, there is still a long way to explore and practice in order for it to be widely applied in learning sciences.