E-learning评论文本的情感分类研究  

Research on Text Sentiment Classification in E-learning

作　　者：潘怡[1] 叶辉[1] 邹军华[2] 

机构地区：[1]长沙学院计算机科学与技术系,湖南长沙410022 [2]湖北大学教育学院,湖北武汉430000

出　　处：《开放教育研究》2014年第2期88-94,共7页Open Education Research

基　　金：湖南省普通高等学校教学改革研究项目"构建有效激励体系;促进软件工程专业实践教育创新"(411)成果之一

摘　　要：自本世纪初起,E—learning作为一种灵活、丰富、高效的学习方式,被越来越多的学习者接受,而伴随着学习技术的逐步成熟,学习者对E—learning应用的要求也从最初的知识推送提升到能够在讲授者与学习者之间搭建有效的沟通桥梁,将零反馈的封闭式学习变成多反馈的协作学习。E—learning的评论信息隐含了学习者在学习中遇到的问题和建议,从中可挖掘学习者对学习资源及授课者的意见。这对改善教学模式、完善教学支持服务意义重要。现有E—learning系统所提供的海量评论信息中正面评论与负面评论夹杂,给挖掘学习者的真实意见和需求带来困难。本文对文本情感分类过程进行归纳,构建了一种情感分类应用模型,在完成预处理、创建词典、提取情感特征后实现了一个情感分类引擎,并将该引擎与实际系统整合。改进后的系统能够将学习者的评论文本自动分为正面评论、负面评论和中性评论,实际性能及用户体验评价结果表明,新的基于情感单元的情感分类方法能满足E—learning评论文本的情感分类需求。Since the beginning of this century, due to its flexibility, variety and high performance, E-learning has been known and accepted by more and more learners. However, with the fast development of information technology, the original role for E-learning has been enlarged from broadcaster to broadcaster and organizer, the traditional zero feedback exclusive learning has been being replaced by a diversified cooperative learning. The comments in E-learning system have implied various questions and suggestions generated by the learners. If we can dig out feedback from the learners bused on their comment in the system, it will help more people to optimize their E-learning teaching model and its supporting ability, which is very important to improve the E-learning system. Nevertheless, due to the positive affirmations and negative criticisms are both mixed in the huge commentary information, it is hard to find out the true requirement and needs of learners. Therefore, we propose a general introduction and conclusion of sentiment classifwa- tion and an application model of sentiment classification in E-Learning. Based on the model, a sentiment classifica- tion engine is integrated with a real E-Learning system, which can automatically divide the commentary texts into three parts, positive, negative, and neutral. The experiment results and user experiences show that the new sentiment classification method combined with the sentiment units satisfies the requirement of the classification of comment texts in E -Learning.

关 键 词：E-LEARNING 评论文本 情感分类 情感单元 

分 类 号：G434[文化科学—教育学]