学习分析研究的现状与未来发展——2017年学习分析与知识国际会议评析

The Development,Status,and Challenges of Learning Analysis:A Review of the 7th International Conference on Learning Analytics and Knowledge

学习分析和知识国际会议(简称LAK)致力于传播和推动学习分析领域的前沿研究。第七届学习分析与知识国际会议(简称LAK17)于2017年3月在加拿大温哥华举行,来自不同学科背景的三位专家作了主题报告。本文对会议主题报告和论文进行系统了分析,以明晰该领域研究发展现状和未来趋势。首先,结合七届LAK会议主题发展,本文从时间维度纵向梳理学习分析的发展历程。然后,围绕本次会议,本文横向探究了学习分析追踪学习、理解学习和改进学习的发展过程,以及学习分析领域可持续发展,具体包括:1)利用多模态生物识别技术采集学习者行为层、心理层和生理层数据,利用线上和线下一体化学习分析技术全方位追踪学习;2)通过学习者行为建模、在线学习社区和反思写作中的文本挖掘、真实课堂环境情境中的师生话语分析、心理测量和情感分析方法,对学习者和学习过程进行多角度理解;3)通过学习者知识建模、学习资源精准服务和自我调节学习支架为学习者提供个性化与自适应学习支持,基于学习分析系统从学习设计和教学决策、学习风险预测和干预、学习评价和反馈等方面为教学者、学习者、研究者和管理者提供多方位教学支持;4)关注学习分析研究的数据权限和伦理道德问题,保证学习分析领域的可持续发展。最后,本文提出了构建多模态数据基础、寻求跨学科技术支持两大跨学科性发展趋势,并指出推动学习分析标准研制和学科建设面临的挑战,以期为学习分析领域的研究发展和以改进学习为目的的实践应用提供启示。

The International Conference on Learning Analytics and Knowledge （LAK） is dedicated to the dis- semination and promotion of cutting-edge research in the field of learning analysis. The 7th International Conference on Learning Analysis and Knowledge （ LAK＇ 17 ） was held in Vancouver, Canada, in March 2017. The theme of LAK＇ 17 was purposely focused on the transdisciplinary nature of research in learning analytics. Three experts from dif- ferent backgrounds shared their keynote speeches. This paper reviews and analyzes the LAK＇ 17 keynotes, articles and posters systematically, exploring the process of tracking learning, understanding learning and improving learning, as well as the sustainability of learning analysis. This part includes the following four aspects ： 1 ） tracking learning by multimodal data analysis, including the multi- modal biometric technology and the online ＆ offlinc consistency analysis of learners%ehavior, psychological ＆ physio- logical data. 2 ） Understanding learning using multi-angle analysis of learners and the learning process, including learner behavior modeling, text mining, discourse analysis, psychological measurement and emotional analysis. 3 ） Improving learning by providing multi-faceted teaching and learning support, including personalized and adaptive learning support （such as learner knowledge modeling, learning resource management and self-regulated learning scaffolding） , learning design and teaching decision-making support, learning risk prediction and intervention, and learning evaluation and feedback etc. 4） The sustainable development of learning analysis, including the discussions on data authority and ethics. Finally, based on the above analyses, this paper puts forward two major trends in the transdisciplinary nature of research in learning analytics ： the construction of multimodal data base, and the search for interdisciplinary technical support. At the same time, this paper points out the challenges in promoting the development of learning and analysis standards and the construction of disciplines, so as to provide inspirations for the development of learning analysis research, and the practical applications that improve learning.