学习分析视角下的个性化学习平台研究

Personalized Learning Platform from the Perspective of the Learning Analytics

个性化学习平台是实现“智慧学习冶的前提，而依托大数据的学习分析技术则是构建实用性强的个性化学习平台的重要保障。然而，目前学习分析技术仍处在发展阶段，相关文献表明现有研究主要集中于理论综述和应用分析，多侧重于基于学习管理系统的数据采集与分析，面向基础教育的成熟应用案例稀少。本研究以基础教育中的智慧学习为应用背景，尝试基于学习分析的动态采集、精确分析、可视化反馈等构建个性化学习平台，以实现信息技术支持下的“因材施教冶。文章从个性化学习的实施要素和学习分析的过程出发，探讨了基础教育中基于学习分析实现个性化学习的思路，即在传统的“一对多冶课堂传播环境中采用教师引导的基于分层的群组个性化学习，在“一对一冶的泛在自主学习环境下采用智能Agent伴学的个性化学习，两种方式可无缝切换，并通过层的粒度设置实现统一;进而阐述了平台的框架、学习路径推送策略、功能模块与智能Agent设计;并结合案例介绍了平台的应用情况。数据表明，引入学习分析技术有助于教师量化掌控学情及动态调整教学过程，对学习困难生和优异生的学习动机激发与保持有较显著的积极作用。

With the popularity of mobile devices， cloud computing and other new technologies， the information technology ecological changes of primary and secondary schools are taking place. Creating a smart learning environment is the most direct way to embrace the changes. A personalized learning platform is a prerequisite to achieve “ smart learning”. The technology of learning analytics based on big data is an important solution to build practical personalized learning platform because it provides a blueprint for the realization of personalized learning. In the development stage of learning analytics， the literature survey shows the existing research papers mainly focus on the review and analysis of the application. The typical case is mainly focusing on the dynamic acquisition and analysis of an LMS platform. Related application cases in K-12 education is rare. Meanwhile， due to the complex nature of education itself， the early “intelligent expert systems” experience has shown that teaching which relies entirely on “ expert systems” or a “ machine mentor”has its deficiencies. Currently， when we try to construct a personalized learning platform， “smart” and “practicality” is still a pair of mutual support and mutual restraints associated variables， and the focus of our work is to find the balance of them. For this reason， we adopted a dynamic granularity group approach to realize personalization. We can control the degree of personalization learning by the granularity. When the granularity is larger than one， the teaching mode is group learning. Otherwise， the teaching model is individualized learning. This paper intends to build a personalized learning platform for K -12 education based on learning analytics through dynamic acquisition， accurate analysis， and visual feedback. Firstly， the author analyzes the implementation of elements of personalized learning and the process of learning analytics. They then discuss the ideas to achieve learning personalization based on learning analytics， which includes two approaches: using instructor -guided， layer - based group-customized learning for traditional ＂ one to many＂ classroom environment， and using intelligent agent guided personalized learning for the ＂ one to one＂ self-learning environment. The two approaches can switch seamlessly， and achieve unity by setting the granularity of layer. And then we describe the framework， function module and intelligent agent design of personalized learning platform from the perspective of learning analytics. In the final section， the application of this platform in an elementary school was introduced. The observation and data analysis shows that the platform can help teachers grasp the situation and adjust their teaching processes， and stimulate students爷learning motivation. The most important factor that influences learning analysis is the scientific integrity of the data. In order to expand the LMS data sources， the brain-computer interface technology to collect students' brain waves will be used in a subsequent study. These direct cognitive data will be used to analyze the students' cognitive