基于ALBERT和同义词词林的主观题自动评分方法
Automatic Scoring Method for Subjective Questions Based on ALBERT and Cilin
DOI: 10.12677/CSA.2020.109177, PDF,    科研立项经费支持
作者: 张展鑫, 陈平华, 刘佳荣：广东工业大学计算机学院，广东 广州
关键词: 主观题；自动评分；相似度；ALBERT；同义词词林；Subjective Question； Automatic Scoring； Similarity； ALBERT； Cilin
摘要: 针对具有参考答案的主观题机器自动评分，既要考虑得分点契合度又要考虑文本整体相似度等问题，提出一种结合ALBERT和同义词词林的主观题自动评分方法。先利用ALBERT的Fine-tuning方法计算参考答案和考生答题之间的文本语义相似度；然后经关键词提取操作，利用同义词词林计算两份文本间面向得分点的关键词相似度；最后结合语义相似度和关键词相似度计算综合得分。真实数据集上的对比实验表明，本文的方法在评分准确率方面有明显提高。
Abstract: Aiming at the machine automatic scoring of subjective questions with reference answers, both the fit of the score points and the overall similarity of the text must be considered. An automatic scoring method for subjective questions that combines ALBERT and Cilin is proposed. First use ALBERT’s Fine-tuning method to calculate the textual semantic similarity between the reference answer and the test taker’s answer. Then use the keyword extraction operation to calculate the score-oriented keyword similarity between the two texts using Cilin. Finally, the comprehensive score is calculated by combining semantic similarity and keyword similarity. Comparative experiments on real data sets show that the method in this paper has a significant improvement in scoring accuracy.
文章引用：张展鑫, 陈平华, 刘佳荣. 基于ALBERT和同义词词林的主观题自动评分方法[J]. 计算机科学与应用, 2020, 10(9): 1673-1682. https://doi.org/10.12677/CSA.2020.109177

参考文献

[1] Page, E.B. (1994) Computer Grading of Student Prose, Using Modern Concepts and Software. The Journal of Experi-mental Education, 62, 127-142. [Google Scholar] [CrossRef] 
[2] 高思丹, 袁春风. 语句相似度计算在主观题自动批改技术中的初步应用[J]. 计算机工程与应用, 2004, 40(14): 132-135.
[3] 倪应华. 基于XML自动阅卷算法的设计与实现[J]. 仪器仪表学报, 2005, 26(z1): 704-705.
[4] Callear, D.H., Jerrams-Smith, J. and Soh, V. (2001) CAA of Short Non-MCQ Answers. Proceedings of the 5th International Computer Assisted As-sessmentconference, Loughborough, July 2001, 1-14.
[5] 王逸凡, 李国平. 基于语义相似度及命名实体识别的主观题自动评分方法[J]. 电子测量技术, 2019, 42(2): 84-87.
[6] 罗海蛟, 柯晓华. 基于改进的LDA模型的中文主观题自动评分研究[J]. 计算机科学, 2017, 44(z2): 102-105+128.
[7] 周洲, 侯开虎, 姚洪发, 张慧. 基于TF-IDF及LSI模型的主观题自动评分系统研究[J]. 软件, 2019, 40(2): 158-163.
[8] 张翠翠, 周国祥, 俞磊, 石雷, 王青青. 基于MVC的试卷生成及主观题判卷算法研究[J]. 系统仿真学报, 2020, 32(1): 105-112.
[9] Devlin, J., Chang, M.W., Lee, K. and Toutanova, K. (2018) Bert: Pre-Training of Deep Bidirectional Transformers for Language Under-standing. arXiv:1810.04805.
[10] Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention Is All You Need. 31st Conference on Neural Information Processing Systems (NIPS 2017), Long Beach, December 2017, 5998-6008.
[11] Lan, Z.Z., Chen, M.D., Goodman, S., Gimpel, K., Sharma, P. and Soricut, R (2019) Albert: A Lite Bert for Self-Supervised Learning of Language Representations. arXiv: 1909.11942.
[12] 梅家驹. 同义词词林[M]. 上海: 上海辞书出版社, 1983.
[13] 哈工大信息检索研究中心. 同义词词林扩展版[DB/OL].
https://www.ltp-cloud.com/download, 2005.
[14] 郭小华, 彭琦, 邓涵, 朱新华. 基于边权重的WordNet词语相似度计算[J]. 计算机工程与应用, 2018, 54(1): 172-178.
[15] 彭琦, 朱新华, 陈意山, 孙柳, 李飞. 基于信息内容的词林词语相似度计算[J]. 计算机应用研究, 2018, 35(2): 400-404.
[16] Liu, X, Chen, Q., Deng, C., et al. (2018) LCQMC: A Large-Scale Chinese Question Matching Corpus. Proceedings of the 27th International Conference on Computational Linguistics, Santa Fe, August 2018,1952-1962.