学习分析数据模型及数据处理方法研究

Research on Data Model and Data Processing Methods in Learning Analytics

数据是学习分析的基础，应用恰当的数据处理方法则是学习分析成功的关键。该文抓住学习分析领域这两个关键点，采用演绎与归纳相结合的方法，即在格框架指导下演绎得出学习分析数据模型的基本框架，参考Moodle学习平台中的数据表结构来丰富数据模型中的各个细节，构建了学习分析数据模型；接着以数据模型中的“学习行为”数据表为例，列出了这一数据表中可能出现的数据子集，以实际案例介绍了针对不同数据子集的处理方法。学习分析实践表明，这些处理方法中最为常用的是分类汇总，在分类汇总的基础上可进一步采用聚类、社会网络分析、文本挖掘等方法，使分析不断走向深入。

Data is the basis of learning analytics， and applying data processing methods appropriately is the key factor to success of learning analytics. This paper concentrated upon the data model and data processing methods of learning analytics. Firstly， the author used both deduction and induction methods to construct data model of learning analytics， which included deducing the basic frame of data model， enriching details of data model by inducing data table structures of Moodle LMS. Secondly， the author chose learning behavior table in data model， listed all possible data subsets in this table， and introduced processing methods according to different data subsets with some examples. Learning analytics practices showed that， the most frequently used method is subtotal， and with the basis of subtotal more methods such as clustering， social network analysis， text mining， etc. could be used.