-
摘要:
大型开放式网络课程(MOOC)的出现虽然极大地改变了人们的学习方式,但用户在MOOC平台开展学习的学习情况及完成率预测仍是目前一个重要的技术挑战。针对预测的需求,从用户的学习行为中对用户和课程进行分析,采用长短时记忆机对学习者的学习活动进行建模,采用多头注意力机制对用户和课程之间的交互活动情况进行分析,提出一个基于门控单元的特征融合框架,用于学习情况预测。在公开数据集上的结果表明:所提框架能够提升预测精度,使得MOOC平台能够尽可能早地对用户活动进行干预,从而提升整体的MOOC平台使用体验。
Abstract:Though the massive open online courses (MOOC) have greatly changed the way of learning, properly understanding the user’s behavior and then predication of dropout is one of the most challenging tasks. MOOC have significantly altered the way that people learn, yet one of the most difficult challenges is correctly interpreting user behavior and then predicting dropout. In this research, to improve the dropout prediction performance, we firstly analyzed users and courses from the perspective of activities by using the long short term memory mechanism. In this study, we used the long short term memory mechanism to analyze users and courses from the perspective of activities in order to improve the dropout prediction performance. Afterwards we further proposed a multi-attention based multi-perspective feature enhancement method to investigate the correlated activities among users and courses. Finally, we provided a gated mechanism-based feature integration framework for dropout prediction. The experiment study on the public dataset has shown our framework’s promising potential, thereby making it possible to better investigate the reason beneath these phenomena and improve the overall study experience. The experiment study on the open dataset has demonstrated the promising potential of our framework, allowing us to more thoroughly explore the causes of these events and enhance the learning environment as a whole.
-
Key words:
- massive open online courses /
- prediction framework /
- user /
- content /
- study behavior
-
表 1 分类结果混淆矩阵
Table 1. Confusion table confusion matrix for classification results
实际标签 预测分类 正例数量 反例数量 正例数量 TP FN 反例数量 FP TN 表 2 ACC和AUC总体性能
Table 2. Overall performace in terms of ACC and AUC
方法 ACC/% AUC/% LR 82.23 SVM 82.86 RF 83.11 DNN 85.64 GBDT 85.18 CFIN 84.78 86.40 本文方法 85.09 86.07 表 3 消融实验
Table 3. Ablation test
方法 ACC/% AUC/% 本文方法 85.09 86.07 w/o SAE 84.97 85.96 w/o MFE 84.91 85.99 w/o GCR 84.91 86.00 注:w/o表示消融实验中的without。 -
[1] MARTIN F G. Will massive open online courses change how we teach?[J]. Communications of the ACM, 2012, 55(8): 26-28. doi: 10.1145/2240236.2240246 [2] LYKOURENTZOU I, GIANNOUKOS I, NIKOLOPOULOS V, et al. Dropout prediction in E-learning courses through the combination of machine learning techniques[J]. Computers & Education, 2009, 53(3): 950-965. [3] FISHBEIN M, AJZEN I. Belief, attitude, intention, and behavior: An introduction to theory and research[J]. Journal of Business Venturing, 1977, 5: 177-189. [4] CHEN C, SONNERT G, SADLER P M, et al. Computational thinking and assignment resubmission predict persistence in a computer science MOOC[J]. Journal of Computer Assisted Learning, 2020, 36(5): 581-594. doi: 10.1111/jcal.12427 [5] CROSSLEY S A, PAQUETTE L, DASCALU M, et al. Combining click-stream data with NLP tools to better understand MOOC completion[C]//Proceedings of the 6th International Conference on Learning Analytics & Knowledge. New York: Association for Computing Machinery, 2016: 6-14. [6] GARDNER J, BROOKS C. Dropout model evaluation in MOOCs[C]//Proceedings of the 32nd AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2018: 7906-7912. [7] DALIPI F, IMRAN A S, KASTRATI Z. MOOC dropout prediction using machine learning techniques: Review and research challenges[C]//Proceedings of 2018 IEEE Global Engineering Education Conference. Piscataway: IEEE Press, 2018: 1007-1014. [8] FENG W, TANG J, LIU T X. Understanding DROPOUTs in MOOCs[C]//Proceedings of the 33rd AAAI Conference on Artificial Intelligence. Palo Alto: AAAI Press, 2019: 517-524. [9] ALRAIMI K M, ZO H, CIGANEK A P. Understanding the MOOCs continuance: The role of openness and reputation[J]. Computers & Education, 2015, 80: 28-38. [10] ITANI A, BRISSON L, GARLATTI S. Understanding learner’s drop-out in MOOCs[C]//Proceedings of 19th International Conference on Intelligent Data Engineering and Automated Learning. Berlin: Springer, 2018: 233-244. [11] YANG D, WEN M, HOWLEY I K, et al. Exploring the effect of confusion in discussion forums of massive open online courses[C]//Proceedings of the 2nd ACM Conference on Learning @ Scale. New York: Association for Computing Machinery, 2015: 121-130. [12] YANG D, SINHA T, ADAMSON D, et al. “Turn on, tune in, drop out”: Anticipating student dropouts in massive open online courses[C]//Proceedings of 2013 NIPS Workshop on Data Driven Education. New York: Curran Associates, 2013. [13] JEON B, PARK N. Dropout prediction over weeks in MOOCs by learning representations of clicks and videos[C]//Proceedings of 2020 AAAI Workshop on Artificial Intelligence for Education. Palo Alto: AAAI Press, 2020. [14] JIN C. MOOC student dropout prediction model based on learning behavior features and parameter optimization[J/OL]. Interactive Learning Environments, 2020 (2020-08-12)[2021-02-12].https://doi.org/10.1080/10494820.2020.1802300. [15] ADAMOPOULOS P. What makes a great MOOC? An interdisciplinary analysis of student retention in online courses[C]//Proceedings of the 2013 International Conference of Information Systems. Milano: AIS eLibrary, 2013. [16] GOOPIO J, CHEUNG C. The MOOC dropout phenomenon and retention strategies[J]. Journal of Teaching in Travel & Tourism, 2021, 21(2): 177-197. [17] XING W, DU D. Dropout prediction in MOOCs: Using deep learning for personalized intervention[J]. Journal of Educational Computing Research, 2019, 57(3): 547-570. doi: 10.1177/0735633118757015 [18] QIU J, TANG J, LIU T X, et al. Modeling and predicting learning behavior in MOOCs[C]//Proceedings of the 9th ACM International Conference on Web Search and Data Mining. New York: Association for Computing Machinery, 2016: 93-102. [19] QIU L, LIU Y, HU Q, et al. Student dropout prediction in massive open online courses by convolutional neural networks[J]. Soft Computing, 2019, 23(20): 10287-10301. doi: 10.1007/s00500-018-3581-3 [20] JEON B, SHAFRAN E, BREITFELLER L, et al. Time-series insights into the process of passing or failing online university courses using neural-induced interpretable student states[C]//Proceedings of the 12th International Conference on Educational Data Mining. Montreal: IEDM, 2019. [21] SINHA T, JERMANN P, LI N, et al. Your click decides your fate: Leveraging clickstream patterns in MOOC videos to infer students’ information processing and attrition behavior[EB/OL]. (2014-09-16)[2021-03-01].https://arxiv.org/abs/1407.7131. [22] YU C H, WU J, LIU A C. Predicting learning outcomes with MOOC clickstreams[J]. Education Sciences, 2014, 9(2): 104. [23] WEN Y, TIAN Y, WEN B, et al. Consideration of the local correlation of learning behavior to predict dropouts from MOOCs[J]. Tsinghua Science and Technology, 2020, 25(3): 336-347. [24] WHITEHILL J, WILLIAMS J J, LOPEZ G, et al. Beyond prediction: Towards automatic intervention in MOOC student stop-out[C]//Proceedings of the 8th International Conference on Educational Data Mining. Madrid: IEDM, 2015: 171-178. [25] FEI M, YEUNG D. Temporal models for predicting student dropout in massive open online courses[C]//Proceedings of 2015 ICDM Workshop on Data Mining for Educational Assessment and Feedback. Piscataway: IEEE Press, 2015: 256-263. [26] WANG W, YU H, MIAO C. Deep model for dropout prediction in MOOCs[C]//Proceedings of the 2nd International Conference on Crowd Science and Engineering. New York: Association for Computing Machinery, 2017: 26-32. [27] TANG C, OUYANG Y, RONG W, et al. Time series model for predicting dropout in massive open online courses[C]//Proceedings of 19th International Conference on Artificial Intelligence in Education. Berlin: Springer, 2018: 353-357. [28] XIONG F, ZOU K, LIU Z, et al. Predicting learning status in MOOCs using LSTM[C]//Proceedings of the 2019 ACM Turing Celebration Conference-China. New York: Association for Computing Machinery, 2019: 1-5. [29] YIN S, LEI L, WANG H, et al. Power of attention in MOOC dropout prediction[J]. IEEE Access, 2020, 8: 202993-203002. doi: 10.1109/ACCESS.2020.3035687 [30] SUN D, MAO Y, DU J, et al. Deep learning for dropout prediction in MOOCs[C]//Proceedings of the 8th International Conference on Educational Innovation through Technology. Piscataway: IEEE Press, 2019. [31] ZHENG Y, GAO Z, WANG Y, et al. MOOC dropout prediction using FWTS-CNN model based on fused feature weighting and time series[J]. IEEE Access, 2020, 8: 225324-225335. doi: 10.1109/ACCESS.2020.3045157 [32] MORENO-MARCOS P M, MUÑOZ-MERINO P J, MALDONADO-MAHAUAD J, et al. Temporal analysis for dropout prediction using self-regulated learning strategies in self-paced MOOCs[J]. Computers & Education, 2020, 145: 103728. [33] KLOFT M, STIEHLER F, ZHENG Z, et al. Predicting MOOC dropout over weeks using machin learning methods[C]//Proceedings of the EMNLP 2014 Workshop on Analysis of Large Scale Social Interaction in MOOCs. Stroudsburg: ACL, 2014: 60-65. [34] NAGRECHA S, DILLON J Z, CHAWLA N V. MOOC dropout prediction: Lessons learned from making pipelines interpretable[C]//Proceedings of the 26th International Conference on World Wide Web Companion. New York: Association for Computing Machinery, 2017: 351-359. [35] LIANG J, LI C, ZHENG L. Machine learning application in MOOCs: Dropout prediction[C]//Proceedings of 11th International Conference on Computer Science & Education. Piscataway: IEEE Press, 2016: 52-57. [36] AL-SHABANDAR R, HUSSAIN A, LAWS A, et al. Machine learning approaches to predict learning outcomes in massive open online courses[C]//Proceedings of 2017 International Joint Conference on Neural Networks. Piscataway: IEEE Press, 2017: 713-720. [37] HANLEY J A, MCNEIL B J. The meaning and use of the area under a receiver operating characteristic (ROC) curve[J]. Radiology, 1982, 143(1): 29-36. doi: 10.1148/radiology.143.1.7063747