信息抽取技术在移动学习资源建设中的应用研究

Research on Information Extraction in Resources Construction of Mobile Learning

在网络教学资源极为丰富的今天,从网页中自动抽取多媒体类及文本类教学资源切片,建立移动学习资源库,既充分提高了现有资源的利用率,也为现有电子书包等移动学习终端提供了丰富的资源来源。文章结合移动学习微内容设计要求,根据多媒体HTML标签特点,及文本类资源递归表达特征,提出了基于信息抽取的移动学习资源片段自动抽取的技术方案,该方案可以实时自动地从Web页面中抽取图片、音视频等多媒体资源切片及填空、选择题等移动学习资源切片。实验结果表明,方案整体准确率达85.4%,召回率达79.0%,时间性能上,可以在113.9小时内获取186,133个移动学习资源切片。该自动化的移动学习资源抽取技术方案具有较高的实用性,也是现有大数据技术在资源建设方面的应用之一。

With abundant educational Web resources, to extract multimedia and text resource clips automatically and to establish a mobile learning resource bank both improve the usability of existing resources and provide rich resources for mobile learning terminals such as e-schoolbags. Combined with the design requirement of micro content of mobile learning, according to the characteristics of muhimedia HTML tags and text resources recursive expression, this paper proposes a technical solution for automatic extraction of mobile learning resources based on information extraction. This solution can automatically extract multimedia resource clips such as images, audio and video, and other mobile learning resources sections such as filling in the blanks, multiple choices from Web pages in real time. The experimental results show that the overall accuracy of the solution is $5.4%, and the recall rate is 79.0%. In terms of time performance,186,133 mobile learning resource clips can be obtained in 113.9 hours. The solution has a high practicability and is one of the applications of big data in the construction of resources.