HU Rong,CUI Rongyi,ZHAO Yahui*.Sentiment analysis of curriculum review based on sentiment dictionary[J].Journal of Yanbian University,2019,45(02):153-160.
基于情感词典的课程评论情感分析
- Title:
- Sentiment analysis of curriculum review based on sentiment dictionary
- 文章编号:
- 1004-4353(2019)02-0153-08
- 分类号:
- TP391
- 文献标志码:
- A
- 摘要:
- 采用极性计算方法,对MOOCs上的课程评论进行情感分析.首先,从MOOCs上搜集课程评论,并对所有评论按学习者、授课方式、课件、平台和视频这5个属性进行分类.其次,基于情感词典对各属性评论进行极性计算,得到各属性的正向评论、负向评论、中性评论和无效评论,将正向评论和负向评论作用于课程评分.最后,分别计算课程评论与5个属性在网页中的共现频率,并将单个共现频率与总共现频率之比作为评分的权重,从教育者角度、学习者角度和平台角度对课程进行评分.将本文方法应用于某高校的课程分析中,结果表明本文方法所得的结果客观、合理.
- Abstract:
- A method of polarity calculation was used to analyze the sentiment of curriculum reviews on MOOCs. First, course reviews were collected from MOOCs, and all reviews were classified according to five attributes: learners, teaching methods, courseware, platform and video. Then, based on the sentiment dictionary, we calculated the polarity of each attribute comment, and got the positive, negative, neutral and invalid comment of each attribute. The positive and negative comment acted on the course score. Finally, the co-occurrence frequencies of curriculum reviews and five attributes in web pages were calculated, and the ratio of single co-occurrence frequency to total co-occurrence frequency was taken as the weight of the scoring. The curriculum was scored from the perspectives of educators, learners and platforms. After applying this method to the educational administration department of a university, the analysis shows that the results obtained by this method are objective and reasonable.
参考文献/References:
[1] 刘兵.情感分析:挖掘观点、情感和情绪[M].刘康,赵军,译.北京:机械工业出版社,2017:1-2.
[2] 林斌.基于语义技术的中文信息情感分析方法研究[D].哈尔滨:哈尔滨工业大学,2006.
[3] PEI J, ZHANG C, HUANG D, et al. Combining word embedding and semantic lexicon for Chinese word similarity computation[C]//Natural Language Understanding and Intelligence Applications. Springer International Publishing, 2016:766-777.
[4] 张磊.细粒度意见挖掘研究[D].南京:南京师范大学,2016.
[5] 杨伊态.在线评论细粒度属性情感分类模型研究[D].武汉:武汉纺织大学,2018.
[6] 刘磊,李学勇,黄永峰.基于集成学习的情感模糊计算分类方法[J].计算机工程与设计,2018:39(7):1872-1876.
[7] 李志民.信息技术发展与教育变革[J].中国教育,2015(9):39-43.
[8] 冯君.基于条件随机场的情感分析模型在MOOCs评论文本分析中的应用研究[D].武汉:华中师范大学,2017.
[9] 陈雷.面向股票评论的情感分析系统研究与实现[D].杭州:浙江工商大学,2017.
[10] 周练.Word2vec的工作原理及应用探究[J].科技情报开发与经济,2015(2):145-148.
[11] 刘娇.基于深度学习的多语种短文本分类方法的研究[D].延吉:延边大学,2018.
[12] 王儒.基于卷积神经网络的短文本表示与分类研究[D].济南:山东师范大学,2018.
[13] 施寒潇.细粒度情感分析研究[D].苏州:苏州大学,2013.
[14] 朱军,刘嘉勇,张腾飞,等.基于情感词典和集成学习的情感极性分类方法[J].计算机应用,2018(38):95-98.
[15] 王科,夏睿.情感词典自动构建方法综述[J].自动化学报,2016:42(4):495-511.
[16] 吴金源,冀俊忠,赵学武,等.基于特征选择技术的情感词权重计算[J].北京工业大学学报,2016(42):142-151.
[17] 徐健.术语相似度计算方法研究[M].广州:中山大学出版社,2012:83-97.表4 属性分类结果
方法 正确率/%本文方法 89.9
备注/Memo
收稿日期: 2019-01-13 *通信作者: 赵亚慧(1974—),女,副教授,研究方向为自然语言文本处理.
*基金项目: 吉林省教育厅职业教育与成人教育教学改革研究课题(2018ZCY334)