智能时代的教育文本挖掘模型与应用  

Educational Text Mining Model and Its Application in the Age of Intelligence

作　　者：刘清堂 贺黎鸣 吴林静 杨炜钦 李晶 LIU Qingtang;HE Liming;WU Linjing;YANG Weiqin;LI Jing

机构地区：[1]华中师范大学教育信息技术学院,湖北武汉430079 [2]华中师范大学教育信息技术协同创新中心

出　　处：《现代远程教育研究》2020年第5期95-103,共9页Modern Distance Education Research

基　　金：国家自然科学基金项目“网络学习资源深度聚合及个性化服务机制研究”(71704062);华中师范大学基本科研业务费专项资金项目“大数据驱动的学习者认知过程建模研究”(CCNU20ZN003)。

摘　　要：教育文本挖掘是指通过数据采集和处理,利用数据挖掘算法或工具,从非结构化文本文档中提取有意义的模式或知识的过程。教育文本蕴含着丰富的学习者认知、行为和情感等信息,对其进行深度挖掘和分析,有助于深入探索教育教学的基本规律,解释教育中存在的问题和现象。大数据支持下的教育文本挖掘模型包括数据生产和使用的主体(利益相关者)、教学环境、数据和挖掘工具等核心要素,涉及数据产生、数据采集、数据处理、知识发现、评估解释、教学应用等过程和方法。其常用的数据来源包含问卷调查、在线互动、学习反馈、在线评论、社交媒体和教学文件,主要用于学习者成绩预测、学习者建模、学习者水平评价、教学材料结构分析、学习者反馈和内容可视化等。当前教育文本挖掘在海量数据处理、数据降维保真、结果评估与解释等方面还面临挑战,研究者需深度融合教育学、认知心理学、语言学等多学科研究方法,结合教育教学的基本理论和具体的教育情境,注重多模态分析和验证,保证将其应用于教育研究的科学性。随着相关技术的突破和应用发展,教育文本数据将成为教育现代化发展的推动力,在深度学习、精准教学等领域中发挥更大作用。Educational text mining refers to the process of using data mining algorithms or tools to extract meaningful patterns or knowledge from unstructured text documents through data collection and processing.Educational texts contain a wealth of information about learners’cognition,behaviors and emotions.Deep mining and analysis of them will help to explore the basic laws of education and teaching,and explain the problems and phenomena in education.The educational text mining model supported by big data includes the main body of data production and use(stakeholders),teaching environment,data and mining tools and other core elements,involving data generation,data collection,data processing,knowledge discovery,evaluation and interpretation,and instructional application and other processes and methods.Its commonly used data sources include questionnaires,online interactions,learning feedback,online comments,social media and teaching documents,which are mainly used for learners’performance prediction,learner modeling,learners’level evaluation,teaching material structure analysis,and learners’feedback and content visualization.Currently,educational text mining still faces challenges in massive data processing,dimensionality reduction and fidelity of data,result evaluation and interpretation,etc.Researchers need to deeply integrate pedagogy,cognitive psychology,linguistics and other multidisciplinary research methods,combined with basic theories of education and teaching and specific educational situations,pay attention to multimodal analysis and verification,to ensure the scientific nature of its application in educational research.With the breakthroughs and application development of related technologies,educational text data will become the driving force for the modernization of education and play a greater role in the fields of deep learning and precision teaching.

关 键 词：教育大数据 数据挖据 学习分析 教育文本挖掘 

分 类 号：G434[文化科学—教育学]