National University of Defense Technology, Changsha410072, China.
Show Author Information
Hide Author Information
Abstract
As a supplement to traditional education, online courses offer people, regardless of their age, gender, or profession, the chance to access state-of-the-art knowledge. Nonetheless, despite the large number of students who choose to begin online courses, it is easy to observe that quite a few of them drop out in the middle, and information on this is vital for course organizers to improve their curriculum outlines. In this work, in order to make a precise prediction of the drop-out rate, we propose a combined method MOOP, which consists of a global tensor and local tensor to express all available feature aspects. Specifically, the global tensor structure is proposed to model the data of the online courses, while a local tensor is clustered to capture the inner connection of courses. Consequently, drop-out prediction is achieved by adopting a high-accuracy low-rank tensor completion method, equipped with a pigeon-inspired algorithm to optimize the parameters. The proposed method is empirically evaluated on real-world Massive Open Online Courses (MOOC) data, and is demonstrated to offer remarkable superiority over alternatives in terms of efficiency and accuracy.
No abstract is available for this article. Click the button above to view the PDF directly.
References
[1]
C.,Alario-HoyosM.,Pérez-SanagustínC. D.Kloos, and P. J.Muñoz-Merino, Recommendations for the design and deployment of MOOCs: Insights about the MOOC digital education of the future deployed in MiríadaX, in Proc. 2nd Int. Conf. Technological Ecosystems for Enhancing Multiculturality, Salamanca, Spain, 2014, pp. 403-408.
[2]
H.Khalil and M.Ebner, MOOCs completion rates and possible methods to improve retention—A literature review, in Proc. World Conf. Educational Multimedia, Hypermedia and Telecommunications, Chesapeake, VA, USA, 2014, pp. 1236-1244.
[3]
D. F. O.,OnahJ.Sinclair, and R.Boyatt, Dropout rates of massive open online courses: Behavioural patterns, in Proc. 6th Int. Conf. Education and New Learning Technologies, Barcelona, Spain, 2014, pp. 5825-5834.
[4]
J. Z.,QiuJ.,TangT. X.,LiuJ.,GongC. H.,ZhangQ.Zhang, and Y. F.Xue, Modeling and predicting learning behavior in MOOCs, in Proc. 9th ACM Int. Conf. Web Search and Data Mining, San Francisco, CA, USA, 2016, pp. 93-102.
[5]
G. L.,ZhangT. J.,AndersonM. W.Ohland, and B. R.Thorndyke, Identifying factors influencing engineering student graduation: A longitudinal and cross-institutional study, J. Eng. Educ., vol. 93, no. 4, pp. 313-320, 2004.
G.,ThomasA.Henderson, and T. L.Goldfinch, The influence of university entry scores on student performance in engineering mechanics, in Proc. Australasian Association for Engineering Education Conf., Adelaide, Australia, 2009, pp. 980-985.
[7]
M.Sheard, Hardiness commitment, gender, and age differentiate university academic performance, Br. J. Educ. Psychol., vol. 79, no. 1, pp. 189-204, 2009.
B. F.,FrenchJ. C.Immekus, and W. C.Oakes, An examination of indicators of engineering students’ success and persistence, J. Eng. Educ., vol. 94, no. 4, pp. 419-425, 2005.
A. Y.Wang and M. H.Newlin, Predictors of web-student performance: The role of self-efficacy and reasons for taking an on-line class, Comput. Hum. Behav., vol. 18, no. 2, pp. 151-163, 2002.
N.Elouazizi, Point-of-view mining and cognitive presence in MOOCs: A (computational) linguistics perspective, in Proc. 2014 Conf. Empirical Methods in Natural Language Processing, Doha, Qatar, 2014, pp. 32-37.
S.,MoonS.Potdar, and L.Martin, Identifying student leaders from MOOC discussion forums through language influence, in Proc. 2014 Conf. Empirical Methods in Natural Language Processing, Doha, Qatar, 2014, pp. 15-20.
S.,CrossleyD. S.,McNamaraR.,BakerY.,WangL.,PaquetteT.Barnes, and Y.Bergner, Language to completion: Success in an educational data mining massive open online class, in Proc. Int. Conf. Educational Data Mining, Madrid, Spain, 2015.
[14]
M. M.,WenD. Y.Yang, and C. P.Rosé, Sentiment analysis in MOOC discussion forums: What does it tell us? in Proc. of Educational Data Mining, London, UK, 2014.
[15]
M. M.,WenD. Y.Yang, and C. P.Rosé, Linguistic reflections of student engagement in massive open online courses, in Proc. 8th Int. AAAI Conf. Web and Social Media, Ann Arbor, MI, USA, 2014.
[16]
D.,ChaplotE.Rhim, and J.Kim, Predicting student attrition in MOOCs using sentiment analysis and neural networks, in Proc. 17th Int. Conf. Artificial Intelligence in Education, Madrid, Spain, 2015.
[17]
J.,WhitehillK.,MohanD.,SeatonY.Rosen, and D.Tingley, Delving deeper into MOOC student dropout prediction, arXiv preprint arXiv:1702.06404, 2017.
T.,SinhaP.,JermannN.Li, and P.Dillenbourg, Your click decides your fate: Inferring information processing and attrition behavior from MOOC video clickstream interactions, in Proc. 2014 Conf. on Empirical Methods in Natural Language Processing, Doha, Qatar, 2014.
[19]
T.,SinhaN.,LiP.Jermann, and P.Dillenbourg, Capturing "attrition intensifying" structural traits from didactic interaction sequences of MOOC learners, in Proc. 2014 Conf. Empirical Methods in Natural Language Processing, Doha, Qatar, 2014, pp. 42-49.
[20]
M.,KloftF.,StiehlerZ. L.Zheng, and N.Pinkwart, Predicting MOOC dropout over weeks using machine learning methods, in Proc. 2014 Conf. Empirical Methods in Natural Language Processing, Doha, Qatar, 2014, pp. 60-65.
[21]
S.,NagrechaJ. Z.Dillon, and N. V.Chawla, MOOC dropout prediction: Lessons learned from making pipelines interpretable, in Proc. 26th Int. Conf. World Wide Web Companion, Perth, Australia, 2017, pp. 351-359.
[22]
G. K.Balakrishnan and D.Coetzee, Predicting Student Retention in Massive Open Online Courses Using Hidden Markov Models, University of California, Berkeley, CA, USA, Tech. Rep. No. UCB/EEC5-2013-109, 2013.
[23]
R. F.Kizilcec and S.Halawa, Attrition and achievement gaps in online learning, in Proc. 2nd ACM Conf. Learning, Vancouver, Canada, 2015, pp. 57-66.
C.,RobinsonM.,YeomansJ.,ReichC.Hulleman, and H.Gehlbach, Forecasting student achievement in MOOCs with natural language processing, in Proc. 6th Int. Conf. Learning Analytics & Knowledge, Edinburgh, UK, 2016, pp. 383-387.
C.Ye and G.Biswas, Early prediction of student dropout and performance in MOOCSs using higher granularity temporal information, J. Learn. Anal., vol. 1, no. 3, pp. 169-172, 2014.
H. B.Duan and P. X.Qiao, Pigeon-inspired optimization: A new swarm intelligence optimizer for air robot path planning, Int. J. Intell. Comput. Cybern., vol. 7, no. 1, pp. 24-37, 2014.
10.26599/TST.2018.9010110.F1
Sketch of course drop-out problem.
3.1 Tensor structure for prediction
We first throw light on the basics of tensors, and then present the tensor structure equipped with a fast low-rank tensor completion for drop-out prediction.
Tensor basics. A tensor is a high-dimensional data representation, whose expression is vector (1-mode) and matrix (2-mode). An -mode tensor can be defined as , where denotes the quantity of mode , and its elements are denoted as , where . The matriculating operator, which unfolds a tensor into a matrix, is defined as , in which the tensor element is mapped to the matrix element , where
The reverse of the matriculation is defined as in a similar way.
The inner product of two same-size tensors is defined as the sum of the products of their entries:
For any , the product of a matrix with a tensor is expressed as , and transformed into the product of two matrixes.
Denote as the Frobenius norm of a tensor. It is clear that
Global tensor. Suppose that there are courses and students in the MOOC data. In order to employ the excellent explanatory power of the tensor structure, we construct the data into 3-dimension tensor , as
Fig. 2
shows.
10.26599/TST.2018.9010110.F2
Sketch of global tensor.
The abscissa axis and vertical axis represent the students, which means the scale of each matrix in the tensor is , and the other dimension represents the course. Thus the size of is , and we randomly select a slice of size . There are different value interpretations for : (1) The value represents "student does not select the course " ; (2) the value represents "student does select the course , and fails to finish it" ; (3) the value represents "student does select the course , and finishes it" ; (4) the value represents "student and student do not select the same course " ; (5) the value represents "student and student do select the same course , and they both fail to finish it" ; and (6) the value represents "student and student select the same course , and they both finish it" . As we can see, the tensor can not only express the drop-out performance of all students in all courses, but also reveal the relationship between students in the specific course. Therefore, we put all data into one tensor to construct a global tensor as a step in generating a MOOP for drop-out prediction.
10.26599/TST.2018.9010110.F2
Sketch of global tensor.
10.26599/TST.2018.9010110.F3
Sketch of local tensor.
10.26599/TST.2018.9010110.F4
Sketch of connecting global and local tensors.
10.26599/TST.2018.9010110.F5
Data statistics. (a) The amount of registering students about each course; (b) the amount of courses each student registering; and (c) detailed statistics of the data.
10.26599/TST.2018.9010110.F6
Variation trend of J-Sim.
10.26599/TST.2018.9010110.F7
Visualization of clutering result.