A new recommendation algorithm combined with spectral clustering and transfer learning
- 153 Downloads
Collaborative filtering (CF) recommendation algorithm has been successfully applied into recommender systems for years which can solve the problem of information overload. However, CF suffers from data sparsity and knowledge cannot transfer between multiple rating matrixes. In this paper, we propose a collaborative filtering recommendation algorithm combined with spectral clustering and transfer learning (RASCTL). RASCTL firstly uses spectral clustering to cluster the dimensions of users and items in the original rating matrix. In addition, RASCTL decomposes the rating matrix after clustering and gets the sharing group rating matrix. Finally, RASCTL makes rating forecasting and recommendations combined with the sharing group rating matrix and transfer learning. By the simulation experiments on Epinions and MovieLents data sets, the results show that RASCTL is able to obtain comparable or even better recommendation accuracy and generalization ability compared with other seven CF recommendation algorithms.
KeywordsSpectral clustering Recommender systems Collaborative filtering Transfer learning
This research was financially supported by University Science Research Project of Jiangsu Province (15KJB520004), Science and Technology Projects of Huaian (HAC201601), Science and Technology Project of Jiangsu Province (BE2015127) and Jiangsu Government Scholarship for Overseas Studies.
Compliance with ethical standards
Conflicts of interest
The authors declare that they have no conflict of interest.
Research involving animal participants
This article does not contain any studies with animals performed by any of the authors.
It was obtained from all individual participants included in the study.
- 2.Guha, S., Mishra, N.: Clustering data streams. In: Minos, G., Johannes, G. (eds.) Data Stream Management, pp. 359–366. Springer, Berlin (2016)Google Scholar
- 16.Li, W.Z.: Fast program for clustering and comparing large sets of protein or nucleotide sequences. In: Karen E.N. (ed.)Encyclopedia of Metagenomics, pp. 173–177. Springer Science+Business Media, New Work (2015). doi: 10.1007/978-1-4899-7478-5_221
- 17.Schroff, F., Kalenichenko, D., Philbin, J.: FaceNet: a unified embedding for face recognition and clustering. In: Computer Vision and Pattern Recognition, Boston, pp. 815–823 (2015). doi: 10.1109/CVPR.2015.7298682
- 18.Bi, C., Wang, H., Bao, R.: SAR image change detection using regularized dictionary learning and fuzzy clustering. In: International Conference on Cloud Computing and Intelligence Systems, Beijing, pp. 327–330 (2015). doi: 10.1109/CCIS.2014.7175753
- 19.Schulam, P., Wigley, F., Saria, S.: Clustering longitudinal clinical marker trajectories from electronic health data: applications to phenotyping and endotype discovery. In: Twenty-Ninth AAAI Conference on Artificial Intelligence, Austin, pp. 2956–2964 (2015)Google Scholar
- 26.Agni, D., Herve, J., Laurent, A.: Image retrieval with reciprocal and shared nearest neighbors. In: 2014 International Conference on Computer Vision Theory and Applications, Lisbon, pp. 321–328 (2014)Google Scholar
- 27.Luiz, P., Tomasz, R., Joshua, A.: Recommending people to people: the nature of reciprocal recommenders with a case study in online dating. User Model. User Adapt. Interact. 23(5), 77–488 (2013)Google Scholar
- 32.Liu, W., Wu, C., Feng, B.: Conditional preference in recommender systems. Expert Syst. Appl. 42(2), 774–788 (2015). doi: 10.1016/j.eswa.2014.08.044
- 36.Gantner, Z., Rendle, S., Freudenthaler, C.: MyMediaLite: a free recommender system library. In: Proceedings of the 15th ACM Conference on Recommender Systems, New York, pp. 305–308 (2011). doi: 10.1145/2043932.2043989