Abstract
Matrix factorization plays a fundamental role in collaborative filtering. There are two basic disciplines among collaborative filtering approaches: neighborhood-based methods and latent factor models. Based on the neighbor-entity spatial relationships, neighborhood-based methods capture the local structure of the user-item rating matrix. Latent factor models capture the global structure of the matrix. Neither neighborhood-based methods nor latent factor models can capture both of them. The recently developed capsule network can capture the part-whole spatial relationships in the images. The basic matrix factorization model and its extensions are among the most successful latent factor models. Motivated by the need for capturing both the local structure and the global structure of the matrix, and inspired by the recently developed capsule network, we propose a new matrix factorization model called capsule matrix factorization, which attempts to capture the two structure of the matrix by propagating the neighbor-entity spatial relationships in the rating matrix into the latent factor vectors. Experimental results on real datasets demonstrate that the capsule matrix factorization model improves the basic matrix model in terms of recommendation accuracy greatly.
Supported by National Science and Technology Supporting Plan No. 2017YFC0804307.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Belkin, M., Niyogi, P.: Laplacian eigenmaps for dimensionality reduction and data representation. Neural Comput. 15, 1373–1396 (2002)
van den Berg, R., Kipf, T.N., Welling, M.: Graph convolutional matrix completion. arXiv abs/1706.02263 (2017)
Chen, C., Li, D., Lv, Q., Yan, J., Chu, S.M., Shang, L.: MPMA: mixture probabilistic matrix approximation for collaborative filtering. In: IJCAI (2016)
Chen, C., Li, D., Lv, Q., Yan, J., Shang, L., Chu, S.M.: GLOMA: embedding global information in local matrix approximation models for collaborative filtering. In: AAAI (2017)
Chen, C., Li, D., Zhao, Y., Lv, Q., Shang, L.: WEMAREC: accurate and scalable recommendation through weighted and ensemble matrix approximation. In: SIGIR (2015)
Dziugaite, G.K., Roy, D.M.: Neural network matrix factorization. arXiv preprint arXiv:1511.06443 (2015)
Hartford, J.S., Graham, D.R., Leyton-Brown, K., Ravanbakhsh, S.: Deep models of interactions across sets. In: ICML (2018)
Jeong, T., Lee, Y., Kim, H.: Ladder capsule network. In: ICML (2019)
Koren, Y.: Factorization meets the neighborhood: a multifaceted collaborative filtering model. In: KDD (2008)
Koren, Y., Bell, R., Volinsky, C.: Matrix factorization techniques for recommender systems. Computer 42(8), 30–37 (2009)
Lee, J., Kim, S., Lebanon, G., Singer, Y.: Local low-rank matrix approximation. In: International Conference on Machine Learning, pp. 82–90 (2013)
sheng Li, D., Chen, C., Liu, W., Lu, T., Gu, N., Chu, S.M.: Mixture-rank matrix approximation for collaborative filtering. In: NIPS (2017)
sheng Li, D., et al.: AdaError: an adaptive learning rate method for matrix approximation-based collaborative filtering. In: WWW (2018)
sheng Li, D., Chen, C., Lv, Q., Shang, L., Chu, S.M., Zha, H.: ERMMA: expected risk minimization for matrix approximation-based recommender systems. In: AAAI (2017)
sheng Li, D., Chen, C., Lv, Q., Yan, J., Shang, L., Chu, S.M.: Low-rank matrix approximation with stability. In: ICML (2016)
Ning, X., Desrosiers, C., Karypis, G.: A comprehensive survey of neighborhood-based recommendation methods. In: Recommender Systems Handbook (2015)
Ning, X., Karypis, G.: SLIM: sparse linear methods for top-n recommender systems. In: 2011 IEEE 11th International Conference on Data Mining, pp. 497–506 (2011)
Omohundro, S.M.: Five balltree construction algorithms (1989)
Paterek, A.: Improving regularized singular value decomposition for collaborative filtering (2007)
Rendle, S., Zhang, L., Koren, Y.: On the difficulty of evaluating baselines: a study on recommender systems. arXiv abs/1905.01395 (2019)
Sabour, S., Frosst, N., Hinton, G.E.: Dynamic routing between capsules. arXiv abs/1710.09829 (2017)
Salakhutdinov, R., Mnih, A.: Bayesian probabilistic matrix factorization using Markov chain Monte Carlo. In: ICML 2008 (2008)
Zhang, J., Shi, X., Zhao, S., King, I.: STAR-GCN: stacked and reconstructed graph convolutional networks for recommender systems. In: Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI 2019, Macao, China, 10–16 August 2019 (2019)
Zheng, Y., Tang, B., Ding, W., Zhou, H.: A neural autoregressive approach to collaborative filtering. In: Proceedings of the 33nd International Conference on Machine Learning, ICML (2016)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Switzerland AG
About this paper
Cite this paper
Li, X., Wu, Y., Zhang, L. (2021). Nonlinear Matrix Factorization via Neighbor Embedding. In: Karlapalem, K., et al. Advances in Knowledge Discovery and Data Mining. PAKDD 2021. Lecture Notes in Computer Science(), vol 12713. Springer, Cham. https://doi.org/10.1007/978-3-030-75765-6_46
Download citation
DOI: https://doi.org/10.1007/978-3-030-75765-6_46
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-75764-9
Online ISBN: 978-3-030-75765-6
eBook Packages: Computer ScienceComputer Science (R0)