Improving the Accuracy of Recommender Systems Through Annealing

  • Shefali Arora
  • Shivani Goel
Conference paper
Part of the Lecture Notes in Networks and Systems book series (LNNS, volume 10)


Matrix factorization is a scalable approach used in recommender systems. It deals with the problem of sparse matrix ratings in datasets. The learning rate parameter in matrix factorization is obtained by using numerical methods like stochastic gradient descent. Learning rate affects the accuracy of the system. In this paper, we make use of annealing schedules which will impact the value of learning rate. Five annealing schedules namely exponential annealing, inverse scaling, logarithmic cooling, linear multiplicative cooling and quadratic multiplicative cooling have been used to affect the learning rate and thus the accuracy of our recommender system. The experimental results on Movielens ( dataset with different sizes show that minimum mean absolute error for the system is obtained by exponential annealing at a lower value of learning rate and by linear multiplicative cooling at higher learning rate values. Apache Mahout (http://www.mahout/ 0.9 is chosen as the platform for conducting the experiments.


Recommender system Matrix factorization Annealing Mahout 


  1. 1.
    Ekstrand, M.D., Riedl, J.T., Konstan J.A.: Collaborative Filtering in Recommender Systems.: Foundations And Trends In Human Computer Interaction, vol. 4, No. 2, pp. 81–173 (2011).Google Scholar
  2. 2.
    Zhang, M., Tang, J., Zhang, X. & Xue X.: Addressing Cold Start in Recommender Systems: A Semi Supervised Co-Training Algorithm. In: Proceedings of the 37th International ACM SIGIR Conference On Research And Development in Information Retrieval, SIGIR’14, pp. 73–82. ACM, New York, NY, USA (2014).Google Scholar
  3. 3.
    Bhebhe, W., Kogeda, O.P.: Shilling Attack Detection in Collaborative Recommender Systems Using a Meta Learning Strategy. In: Emerging Trends In Networks and Computer Communications (ETNCC), International Conference on IEEE, pp. 56–61 (2015).Google Scholar
  4. 4.
    Koren, Y., Bell, R., Volinsky, C.: Matrix Factorization Techniques for Recommender Systems.: Computer, vol. 42, No 8, pp. 30–37 (2009).Google Scholar
  5. 5.
    Luo, X., Xia, Y., Zhu Q.: Applying the Learning rate Adaptation to the Matrix Factorization Based Collaborative Filtering.: Knowledge Based Systems, vol. 37, pp. 154–164 (2013).Google Scholar
  6. 6.
    Chin, W-S., Zhuang, Y., Juan, Y-C., Lin C-J.: A Learning Rate Schedule for Stochastic Gradient Descent Methods to matrix Factorization. In: Advances In Knowledge Discovery And Data Mining, 19th Pacific Asia Conference PAKDD, Ho Chi Minh City, Vietnam, Proceedings, Part I. pp. 442–455 Cham: Springer International Publishing. (2015).Google Scholar
  7. 7.
    Sarwar, B., Karypis, G., Riedl, J.: Application Of Dimensionality Reduction In Recommender System-A Case Study.: ACM WebKDD 2000 Workshop, pp. 1–12. ACM SIGKDD (2000).Google Scholar
  8. 8.
    Le, B.H., Nguyen, K.Q., Thawonmas, R.: Bounded SVD: A Matrix Factorization Method With Bound Constraints For Recommender Systems. In: Emerging Information Technology and Engineering Solutions, pp. 23–26 (2015).Google Scholar
  9. 9.
    Paterek, A.: Improving Regularized Singular Value Decomposition for Collaborative Filtering. In: Proceedings KDD Cup Workshop at SIGKDD’07, 13th ACM International Conference on Knowledge Discovery And Data Mining pp. 39–42. ACM (2007).Google Scholar
  10. 10.
    Yu, H.-F., Hsieh, C.-J., Dhillon I.S.: Scalable Coordinate Descent Approaches to Parallel Matrix Factorization For Recommender Systems. In: IEEE International Conference Of Data mining, pp. 766–774. IEEE (2012).Google Scholar
  11. 11.
    Weimer, M., Karatzoglou, A., Smola, A. Improving Maximum Margin Matrix Factorization. In: Machine Learning, vol. 72, No. 3, pp. 263–276 (2008).Google Scholar
  12. 12.
    Bamnote, G.R., Agrawal S.S. Evaluating and Implementing Collaborative Filtering Systems Using Apache Mahout. In: Computing Communication Control And Automation (ICCUBEA), pp. 858–862 (2015).Google Scholar
  13. 13.
    Andresen, B., Nourani, Y.A.: A Comparison of Simulated Annealing Cooling Strategies: Journal of Physics: Mathematics and General, vol. 31, No. 41, pp. 8373–8380 (2011).Google Scholar
  14. 14.
    Gemulla, R., Nijkamp, E., Hass, P.J, Sismanis, Y.: Large Scale Matrix Factorization With Distributed Stochastic Gradient Descent. In: Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, ser. KDD’11, pp. 69–77. New York, NY, USA (2011).Google Scholar
  15. 15.
    Semainario, C.E., Wilson, D.C.: Case Study Evaluation of Mahout As a Recommender Platform. In: 6th ACM conference on recommender engines, pp. 45–50. ACM (2012).Google Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2018

Authors and Affiliations

  1. 1.Department of Computer Science and EngineeringThapar UniversityPatialaIndia

Personalised recommendations