[1]刘金晶,王丽英.在线学习社区发帖质量评价的回归模型研究[J].南京师范大学学报(工程技术版),2020,(01):033-41.[doi:10.3969/j.issn.1672-1292.2020.01.006]
 Liu Jinjing,Wang Liying.Regression Model Research on Posting Quality Evaluationin Online Learning Community[J].Journal of Nanjing Normal University(Engineering and Technology),2020,(01):033-41.[doi:10.3969/j.issn.1672-1292.2020.01.006]
点击复制

在线学习社区发帖质量评价的回归模型研究
分享到:

南京师范大学学报(工程技术版)[ISSN:1006-6977/CN:61-1281/TN]

卷:
期数:
2020年01期
页码:
033-41
栏目:
信息与通信工程
出版日期:
2020-03-15

文章信息/Info

Title:
Regression Model Research on Posting Quality Evaluationin Online Learning Community
文章编号:
1672-1292(2020)01-0033-09
作者:
刘金晶王丽英
南京师范大学教育科学学院,江苏 南京 210023
Author(s):
Liu JinjingWang Liying
School of Education Science,Nanjing Normal University,Nanjing 210023,China
关键词:
机器学习文本质量评价概念关系图在线学习社区开放话题
Keywords:
machine learningtext quality evaluationconceptual relationship diagramonline learning communityopen topic
分类号:
G434
DOI:
10.3969/j.issn.1672-1292.2020.01.006
文献标志码:
A
摘要:
在线学习社区中,多样化教学情境下基于开放话题的发帖使学生能够阐述自我知识更新的进展,但也伴随着越来越高的阅读评价解析成本. 为此,可应用机器学习理论构建发帖质量评价回归模型来实现文本自动评价. 首先构建文本质量评价指标及其计算所依赖的概念关系图,然后结合专家评分标准与评分结果选用多种拟合回归算法对文本质量进行预测评价,最后以拟合优度、交叉验证精度方差和均方误差为指标评估算法效果模型,测试以倡导知识建构学习理念的“数课”平台《网络安全与维护》课程的575条发帖为数据集,实现了网络安全领域的概念关系图存取和发帖质量的特征提取与预测评价. 实验表明,梯度树上升回归算法的准确性、稳定性均优于其他算法模型. 该回归模型能够从5个与文本质量显著相关的特征维度,即可读性、相关度、内聚度、专业度和探究度,有效地实现文本质量自动评价,从而为教师减负和学生自我诊断提供帮助.
Abstract:
In the online learning community,postings based on open topics in a variety of teaching contexts enable students to articulate the progress of self-knowledge updat, but with the increasing costs of reading, evaluating and analyzing these posts. To alleviate it,the machine learning theory is applied to construct a regression model for posting quality evaluation. Firstly,the model constructs five dimensions of text quality evaluation and the conceptual relationship diagram on which the calculation depends. Secondly,combined with the expert scoring standard and the scoring result,the model uses multiple fitting regression algorithms to predict and evaluate the text quality. Finally,the algorithm is evaluated by coefficient of determination,cross-validation accuracy and mean square error. This paper takes 575 posts of the“Shu Ke”platform which advocates knowledge construction from“Network Security and Maintenance”course as a dataset. The model can realize the feature extraction and predict the posting quality evaluation by means of conceptual relationship diagram. The experimental results show that the accuracy and stability of the gradient decision boosting tree regression algorithm are better than those of other algorithm models. The regression model can effectively realize the automatic evaluation of text quality based on five characteristic dimensions that are significantly related to text quality:readability,relevance,cohesion,professionalism and exploration,thus reducing the burden on teachers and helping students with self-diagnosis.

参考文献/References:

[1] 何玲,黎加厚. 促进学生深度学习[J]. 现代教学,2005(5):29-30.
[2]杨宗凯. 数据驱动个性化学习[N]. 中国教育报,2018-11-01(007).
[3]侯明良. MOOC讨论区数据挖掘与应用[D]. 济南:山东大学,2016.
[4]AU C H,LAM K C,FUNG W S,et al. Using animation to develop a MOOC on information security[C]//IEEE International Conference on Industrial Engineering and Engineering Management.
[5]史文祥. 基于DT-BM的学习者主题行为模型研究[D]. 武汉:华中师范大学,2017.
[6]李敏. 虚拟学习社区成员互动的知识建构效果分析[D]. 扬州:扬州大学,2015.
[7]张平霞. 基于文本挖掘的MOOC讨论区学习评价研究[D]. 重庆:重庆师范大学,2018.
[8]VALENTI S,NEFF E,CUCCHIARELLI A. An overview of current research on automated essay grading[J]. Journal of Information Technology Education,2003,2:319-330.
[9]谭冬晨. 主观题评分算法模型研究[D]. 成都:电子科技大学,2011.
[10]王漪. 文本挖掘技术的研究及其在教学平台中的应用[D]. 北京:北京交通大学,2014.
[11]SHEHAB A,ELHOSENY M,HASSANIEN A E. A hybrid scheme for automated essay grading based on LVQ and NLP techniques[C]//2017 13th International Computer Engineering Conference(ICENCO). Cairo:IEEE,2016:65-70.
[12]RAMALINGAM V V,PANDIAN A,CHETRY P,et al. Automated essay grading using machine learning algorithm[C]//2018 10th National Conference on Mathematical Techniques and its Applications(NCMTA). Kattankulathur,2018.
[13]LIU M,WANG Y Q,XU W W,et al. Automated scoring of Chinese engineering students’ english essays[J]. International Journal of Distance Education Technoloies(IJDET),2017,15(1):52-68.
[14]钟将,张淑芳,郭卫丽,等. 主题特征格分析:一种用户生成文本质量评估方法[J]. 电子学报,2018,46(9):2201-2206.
[15]靳健,季平. 用于在线产品评论质量分析的Co-training算法[J]. 上海大学学报(自然科学版),2014,20(3):289-295.
[16]王洪伟,孟园. 在线评论质量有用特征识别:基于GBDT特征贡献度方法[J]. 中文信息学报,2017,31(3):109-117.
[17]聂卉. 基于内容分析的用户评论质量的评价与预测[J]. 图书情报工作,2014,58(13):83-89.
[18]张艳丰,李贺,彭丽徽,等. 基于模糊神经网络的在线评论效用分类过滤模型研究[J]. 情报科学,2017,35(5):94-99,131.
[19]王忠群,皇苏斌,修宇,等. 基于领域专家和商品特征概念树的在线商品评论深刻性度量[J]. 现代图书情报技术,2015(9):17-25.
[20]AIKA Q,KARIM B S S,RAM G R,et al. A concept-level approach to the analysis of online review helpfulness[J]. Computers in Human Behavior,2016,58:75-81.
[21]NIKOLAOS K,ELENA G B,SALVADOR S A. Evaluating content quality and helpfulness of online product reviews:the interplay of review helpfulness vs. review content[J]. Electronic Commerce Research and Applications,2012,11(3):205-217.
[22]黄传河,杜瑞颖,张沪寅,等. 网络安全[M]. 武汉:武汉大学出版社,2004.
[23]龚越,罗小芹,王殿海,等. 基于梯度提升回归树的城市道路行程时间预测[J]. 浙江大学学报(工学版),2018,52(3):453-460.

相似文献/References:

[1]杨杨,刘会东.一种基于成对约束的特征选择改进算法[J].南京师范大学学报(工程技术版),2011,11(01):056.
 Yang Yang,Liu Huidong.An Improved Algorithm for Feature Selection Based on Pairwise Constraint[J].Journal of Nanjing Normal University(Engineering and Technology),2011,11(01):056.
[2]赵红艳,等.基于机器学习与语义知识的动词隐喻识别[J].南京师范大学学报(工程技术版),2011,11(03):059.
 Zhao Hongyan,Qu Weiguang,et al.Chinese Verb Metaphor Recognition Based on Machine Learning and Semantic Knowledge[J].Journal of Nanjing Normal University(Engineering and Technology),2011,11(01):059.
[3]柏宏权,韩庆年.机器学习在适应性教学系统中的应用研究[J].南京师范大学学报(工程技术版),2007,07(04):076.
 Bai Hongquan,Han Qingnian.Application of Machine Learning in Adaptive Instructional System[J].Journal of Nanjing Normal University(Engineering and Technology),2007,07(01):076.
[4]吴兴惠,吴 迪,周玉萍,等.基于机器学习算法的稀土元素掺杂TiO2光催化活性分析[J].南京师范大学学报(工程技术版),2017,17(03):087.[doi:10.3969/j.issn.1672-1292.2017.03.013]
 Wu Xinghui,Wu Di,Zhou Yuping,et al.Photocatalytic Activity Prediction of Rare Earth Doped TiO2Based on Machine Learning Algorithm[J].Journal of Nanjing Normal University(Engineering and Technology),2017,17(01):087.[doi:10.3969/j.issn.1672-1292.2017.03.013]

备注/Memo

备注/Memo:
收稿日期:2019-05-26.
基金项目:全国教育科学“十三五”规划教育部重点课题(DIA170375).
通讯作者:王丽英,博士,讲师,研究方向:计算机图形学、信息技术. E-mail:wangliying@njnu.edu.cn
更新日期/Last Update: 2020-03-15