To alleviate the problem of data sparsity inherent to recommender systems, we propose a semi-supervised framework for stream-based recommendations. Our framework uses abundant unlabelled information to improve the quality of recommendations. We extend a state-of-the-art matrix factorization algorithm by the ability to add new dimensions to the matrix at runtime and implement two approaches to semi-supervised learning: co-training and self-learning. We introduce a new evaluation protocol including statistical testing and parameter optimization. We then evaluate our framework on five real-world datasets in a stream setting. On all of the datasets our method achieves statistically significant improvements in the quality of recommendations.
%0 Journal Article
%1 Matuszyk2017
%A Matuszyk, Pawel
%A Spiliopoulou, Myra
%D 2017
%J Machine Learning
%K from:matuszyk myown
%P 1--28
%R 10.1007/s10994-016-5614-4
%T Stream-based semi-supervised learning for recommender systems
%U https://kmd.cs.ovgu.de/pub/matuszyk/Stream-based-Semi-supervised-Learning-for-RS.pdf
%X To alleviate the problem of data sparsity inherent to recommender systems, we propose a semi-supervised framework for stream-based recommendations. Our framework uses abundant unlabelled information to improve the quality of recommendations. We extend a state-of-the-art matrix factorization algorithm by the ability to add new dimensions to the matrix at runtime and implement two approaches to semi-supervised learning: co-training and self-learning. We introduce a new evaluation protocol including statistical testing and parameter optimization. We then evaluate our framework on five real-world datasets in a stream setting. On all of the datasets our method achieves statistically significant improvements in the quality of recommendations.
@article{Matuszyk2017,
abstract = {To alleviate the problem of data sparsity inherent to recommender systems, we propose a semi-supervised framework for stream-based recommendations. Our framework uses abundant unlabelled information to improve the quality of recommendations. We extend a state-of-the-art matrix factorization algorithm by the ability to add new dimensions to the matrix at runtime and implement two approaches to semi-supervised learning: co-training and self-learning. We introduce a new evaluation protocol including statistical testing and parameter optimization. We then evaluate our framework on five real-world datasets in a stream setting. On all of the datasets our method achieves statistically significant improvements in the quality of recommendations.},
added-at = {2017-02-08T11:53:04.000+0100},
author = {Matuszyk, Pawel and Spiliopoulou, Myra},
biburl = {https://www.bibsonomy.org/bibtex/2c077e9a7b3f48c3767aa48fb72e5d3db/kmd-ovgu},
doi = {10.1007/s10994-016-5614-4},
interhash = {73baf1e02f8dcebf1a4eca232c58b6bb},
intrahash = {c077e9a7b3f48c3767aa48fb72e5d3db},
issn = {1573-0565},
journal = {Machine Learning},
keywords = {from:matuszyk myown},
pages = {1--28},
timestamp = {2018-03-09T12:46:02.000+0100},
title = {Stream-based semi-supervised learning for recommender systems},
url = {https://kmd.cs.ovgu.de/pub/matuszyk/Stream-based-Semi-supervised-Learning-for-RS.pdf},
year = 2017
}