Enhanced Higher Order Orthogonal Iteration Algorithm for Student Performance Prediction

Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 379)


Predicting Student Performance is the process that predicts the successful completion of a task by a student. Such systems may be modeled using a three-mode tensor where the three entities are user, skill, and task. Recommendation systems have been implemented using Dimensionality reduction techniques like Higher Order Singular Value Decomposition (HOSVD) combined with Kernel smoothing techniques to bring out good results. Higher Order Orthogonal Iteration (HOOI) algorithms have also been used in recommendation systems to bring out the relationship between the three entities, but the prediction results would be largely affected by the sparseness in the tensor model. In this paper, we propose a generic enhancement to HOOI algorithm by combining it with Kernel smoothing techniques. We perform an experimental comparison of the three techniques using an ITS dataset and show that our proposed method improves the prediction for larger datasets.


Recommendation systems HOSVD Higher order orthogonal iteration algorithm Kernel smoothing Tensors 



This work derives direction and inspiration from the Chancellor of Amrita University, Sri Mata Amritanandamayi Devi. We would like to thank Dr. M. Ramachandra Kaimal, Head, Department of Computer Science and Dr. Bhadrachalam Chitturi, Associate Professor, Department of Computer Science, Amrita University for their valuable feedback.


  1. 1.
    Breese, J.S., Heckerman, D., Kadie, C.: Empirical analysis of predictive algorithms for collaborative filtering. In: Proceedings of 14th Conference Uncertainty in Artificial Intelligence, p. 4352. Morgan Kaufmann (1998)Google Scholar
  2. 2.
    Symeonidis, P., Nanopoulos, A., Manolopoulos, Y.: A unified framework for providing recommendations in social tagging systems based on ternary semantic analysis. IEEE Trans. Knowl. Data Eng. 22(2), (2010)Google Scholar
  3. 3.
    Ishteva, M., Absil, P.-A., Van Huffel, S., De Lathauwer, L.: Best Low multilinear rank approximation of higher-order tensors, based on the riemannian trust-region scheme. SIAM. J. Matrix Anal. Appl. 32(1), 115–135Google Scholar
  4. 4.
    Turney, P.D.: Empirical evaluation of four tensor decomposition algorithms. In: Technical Report ERB-1152, NRC-49877, November 12 2007Google Scholar
  5. 5.
    Koren, Y., Bell, R., Volinsky, C.: Matrix factorization techniques for recommender systems. Computer 42(8), 3037 (2009)CrossRefGoogle Scholar
  6. 6.
    Xu, Y., Zhang, L., Liu, W.: Cubic analysis of social bookmarking for personalized recommendation. In: Frontiers of WWW Research and DevelopmentAPWeb 06, pp. 733–738. Springer (2006)Google Scholar
  7. 7.
    Ning, X., Karypis, G.: Multi-task Learning for recommender Systems. In: JMLR: Workshop and Conference Proceedings, vol. 13, pp. 269–284 2nd Asian Conference on Machine Learning (ACML 2010), Tokyo, Japan, Nov. 810 2010Google Scholar
  8. 8.
    Tso-Sutter, K., Marinho, B., Schmidt-Thieme, L.: Tag-aware recommender systems by fusion of collaborative filtering algorithms. In: Proceedings of ACM Symposium Applied Computing (SAC) Conference (2008)Google Scholar
  9. 9.
    Jäschke, R., Marinho, L., Hotho, A., Schmidt-Thieme, L., Stumme, G.: Tag recommendations in social bookmarking systems. In: AI Communications, p. 231247 (2008)Google Scholar
  10. 10.
    Kolda, T.G., Bader, B.W.: Tensor Decompositions Appl. SIAM Rev. 51(3), 455500 (46 p)Google Scholar
  11. 11.
    Ishteva, M., De Lathauwer, L., Absil, P.: Dimensionality reduction for higher-order tensors: algorithms and applications. Int. J. Pure Appl. Math. 42, 337–343 (2008)Google Scholar
  12. 12.
    Luo, D., Huang, H., Ding, C.: Are tensor decomposition solutions unique? In: On the Global Convergence of HOSVD and ParaFac Algorithms. CVPR2009Google Scholar
  13. 13.
    Sun, J.-T., Zeng, H.-J., Liu, H., Lu, Y., Chen, Z.: CubeSVD: a novel approach to personalized web search. In: Proceedings WWW ‘05 Proceedings of the 14th International Conference on World Wide Web, pp. 382–390 ACM New York, NY, USA (2005)Google Scholar
  14. 14.
    Nanopoulos, A., Krohn-Grimberghe, A.: Recommending in social tagging systems based on kernelized multiway analysis. In: Proceedings of the 11th IFCS Biennial Conference and 33rd Annual Conference of the Gesellschaft für Klassifikation e.V., Dresden, March 13–18 2009Google Scholar
  15. 15.
    Bader, B.W., Kolda, T.G.: Efficient MATLAB computations with sparse and factored tensors. SIAM J. Sci. Comput. 30(1), 205231Google Scholar
  16. 16.
    Stamper, J., Niculescu-Mizil, A., Ritter, S., Gordon, G.J., Koedinger, K.R.: Algebra 2005–06. In: Challenge Dataset from KDD Cup 2010 Educational Data Mining Challenge (2010)Google Scholar
  17. 17.
    Bader, B.W., Kolda, T.G., et al.: MATLAB Tensor Toolbox Version 2.5, Available online, January 2012Google Scholar

Copyright information

© Springer India 2016

Authors and Affiliations

  1. 1.Amrita CREATEAmrita Vishwa VidyapeethamKollamIndia
  2. 2.Department of Computer ScienceAmrita Vishwa VidyapeethamKollamIndia

Personalised recommendations