Local Low-Rank Matrix Approximation

Joonseok Lee, Seungyeon Kim, Guy Lebanon, Yoram Singer
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(2):82-90, 2013.

Abstract

Matrix approximation is a common tool in recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show improvements in prediction accuracy over classical approaches for recommendation tasks.

Cite this Paper


BibTeX
@InProceedings{pmlr-v28-lee13, title = {Local Low-Rank Matrix Approximation}, author = {Lee, Joonseok and Kim, Seungyeon and Lebanon, Guy and Singer, Yoram}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {82--90}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {2}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/lee13.pdf}, url = {https://proceedings.mlr.press/v28/lee13.html}, abstract = {Matrix approximation is a common tool in recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show improvements in prediction accuracy over classical approaches for recommendation tasks.} }
Endnote
%0 Conference Paper %T Local Low-Rank Matrix Approximation %A Joonseok Lee %A Seungyeon Kim %A Guy Lebanon %A Yoram Singer %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-lee13 %I PMLR %P 82--90 %U https://proceedings.mlr.press/v28/lee13.html %V 28 %N 2 %X Matrix approximation is a common tool in recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show improvements in prediction accuracy over classical approaches for recommendation tasks.
RIS
TY - CPAPER TI - Local Low-Rank Matrix Approximation AU - Joonseok Lee AU - Seungyeon Kim AU - Guy Lebanon AU - Yoram Singer BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/05/13 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-lee13 PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 2 SP - 82 EP - 90 L1 - http://proceedings.mlr.press/v28/lee13.pdf UR - https://proceedings.mlr.press/v28/lee13.html AB - Matrix approximation is a common tool in recommendation systems, text mining, and computer vision. A prevalent assumption in constructing matrix approximations is that the partially observed matrix is of low-rank. We propose a new matrix approximation model where we assume instead that the matrix is locally of low-rank, leading to a representation of the observed matrix as a weighted sum of low-rank matrices. We analyze the accuracy of the proposed local low-rank modeling. Our experiments show improvements in prediction accuracy over classical approaches for recommendation tasks. ER -
APA
Lee, J., Kim, S., Lebanon, G. & Singer, Y.. (2013). Local Low-Rank Matrix Approximation. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(2):82-90 Available from https://proceedings.mlr.press/v28/lee13.html.

Related Material