How to Apply Random Projections to Nonnegative Matrix Factorization with Missing Entries?. Yahaya, F., Puigt, M., Delmaire, G., & Roussel, G. In 2019 27th European Signal Processing Conference (EUSIPCO), pages 1-5, Sep., 2019.
Paper doi abstract bibtex Random projections belong to the major techniques to process big data and have been successfully applied to Nonnegative Matrix Factorization (NMF). However, they cannot be applied in the case of missing entries in the matrix to factorize, which occurs in many actual problems with large data matrices. In this paper, we thus aim to solve this issue and we propose a novel framework to apply random projections in weighted NMF, where the weight models the confidence in the data (or the absence of confidence in the case of missing data). We experimentally show the proposed framework to significantly speed-up state-of-the-art NMF methods under some mild conditions. In particular, the proposed strategy is particularly efficient when combined with Nesterov gradient or alternating least squares.
@InProceedings{8903036,
author = {F. Yahaya and M. Puigt and G. Delmaire and G. Roussel},
booktitle = {2019 27th European Signal Processing Conference (EUSIPCO)},
title = {How to Apply Random Projections to Nonnegative Matrix Factorization with Missing Entries?},
year = {2019},
pages = {1-5},
abstract = {Random projections belong to the major techniques to process big data and have been successfully applied to Nonnegative Matrix Factorization (NMF). However, they cannot be applied in the case of missing entries in the matrix to factorize, which occurs in many actual problems with large data matrices. In this paper, we thus aim to solve this issue and we propose a novel framework to apply random projections in weighted NMF, where the weight models the confidence in the data (or the absence of confidence in the case of missing data). We experimentally show the proposed framework to significantly speed-up state-of-the-art NMF methods under some mild conditions. In particular, the proposed strategy is particularly efficient when combined with Nesterov gradient or alternating least squares.},
keywords = {Big Data;gradient methods;least squares approximations;matrix decomposition;NMF methods;apply random projections;Nonnegative Matrix Factorization;missing entries;big data;data matrices;weighted NMF;missing data;Economic indicators;Signal processing algorithms;Estimation;Matrix decomposition;Europe;Signal processing;Sparse matrices;Nonnegative matrix factorization;missing data;random projections;low-rank matrix completion;blind source separation;big data},
doi = {10.23919/EUSIPCO.2019.8903036},
issn = {2076-1465},
month = {Sep.},
url = {https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533937.pdf},
}
Downloads: 0
{"_id":"AtbsvKSLEeuWKDzKE","bibbaseid":"yahaya-puigt-delmaire-roussel-howtoapplyrandomprojectionstononnegativematrixfactorizationwithmissingentries-2019","authorIDs":[],"author_short":["Yahaya, F.","Puigt, M.","Delmaire, G.","Roussel, G."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"firstnames":["F."],"propositions":[],"lastnames":["Yahaya"],"suffixes":[]},{"firstnames":["M."],"propositions":[],"lastnames":["Puigt"],"suffixes":[]},{"firstnames":["G."],"propositions":[],"lastnames":["Delmaire"],"suffixes":[]},{"firstnames":["G."],"propositions":[],"lastnames":["Roussel"],"suffixes":[]}],"booktitle":"2019 27th European Signal Processing Conference (EUSIPCO)","title":"How to Apply Random Projections to Nonnegative Matrix Factorization with Missing Entries?","year":"2019","pages":"1-5","abstract":"Random projections belong to the major techniques to process big data and have been successfully applied to Nonnegative Matrix Factorization (NMF). However, they cannot be applied in the case of missing entries in the matrix to factorize, which occurs in many actual problems with large data matrices. In this paper, we thus aim to solve this issue and we propose a novel framework to apply random projections in weighted NMF, where the weight models the confidence in the data (or the absence of confidence in the case of missing data). We experimentally show the proposed framework to significantly speed-up state-of-the-art NMF methods under some mild conditions. In particular, the proposed strategy is particularly efficient when combined with Nesterov gradient or alternating least squares.","keywords":"Big Data;gradient methods;least squares approximations;matrix decomposition;NMF methods;apply random projections;Nonnegative Matrix Factorization;missing entries;big data;data matrices;weighted NMF;missing data;Economic indicators;Signal processing algorithms;Estimation;Matrix decomposition;Europe;Signal processing;Sparse matrices;Nonnegative matrix factorization;missing data;random projections;low-rank matrix completion;blind source separation;big data","doi":"10.23919/EUSIPCO.2019.8903036","issn":"2076-1465","month":"Sep.","url":"https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533937.pdf","bibtex":"@InProceedings{8903036,\n author = {F. Yahaya and M. Puigt and G. Delmaire and G. Roussel},\n booktitle = {2019 27th European Signal Processing Conference (EUSIPCO)},\n title = {How to Apply Random Projections to Nonnegative Matrix Factorization with Missing Entries?},\n year = {2019},\n pages = {1-5},\n abstract = {Random projections belong to the major techniques to process big data and have been successfully applied to Nonnegative Matrix Factorization (NMF). However, they cannot be applied in the case of missing entries in the matrix to factorize, which occurs in many actual problems with large data matrices. In this paper, we thus aim to solve this issue and we propose a novel framework to apply random projections in weighted NMF, where the weight models the confidence in the data (or the absence of confidence in the case of missing data). We experimentally show the proposed framework to significantly speed-up state-of-the-art NMF methods under some mild conditions. In particular, the proposed strategy is particularly efficient when combined with Nesterov gradient or alternating least squares.},\n keywords = {Big Data;gradient methods;least squares approximations;matrix decomposition;NMF methods;apply random projections;Nonnegative Matrix Factorization;missing entries;big data;data matrices;weighted NMF;missing data;Economic indicators;Signal processing algorithms;Estimation;Matrix decomposition;Europe;Signal processing;Sparse matrices;Nonnegative matrix factorization;missing data;random projections;low-rank matrix completion;blind source separation;big data},\n doi = {10.23919/EUSIPCO.2019.8903036},\n issn = {2076-1465},\n month = {Sep.},\n url = {https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533937.pdf},\n}\n\n","author_short":["Yahaya, F.","Puigt, M.","Delmaire, G.","Roussel, G."],"key":"8903036","id":"8903036","bibbaseid":"yahaya-puigt-delmaire-roussel-howtoapplyrandomprojectionstononnegativematrixfactorizationwithmissingentries-2019","role":"author","urls":{"Paper":"https://www.eurasip.org/proceedings/eusipco/eusipco2019/proceedings/papers/1570533937.pdf"},"keyword":["Big Data;gradient methods;least squares approximations;matrix decomposition;NMF methods;apply random projections;Nonnegative Matrix Factorization;missing entries;big data;data matrices;weighted NMF;missing data;Economic indicators;Signal processing algorithms;Estimation;Matrix decomposition;Europe;Signal processing;Sparse matrices;Nonnegative matrix factorization;missing data;random projections;low-rank matrix completion;blind source separation;big data"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"inproceedings","biburl":"https://raw.githubusercontent.com/Roznn/EUSIPCO/main/eusipco2019url.bib","creationDate":"2021-02-11T19:15:22.104Z","downloads":0,"keywords":["big data;gradient methods;least squares approximations;matrix decomposition;nmf methods;apply random projections;nonnegative matrix factorization;missing entries;big data;data matrices;weighted nmf;missing data;economic indicators;signal processing algorithms;estimation;matrix decomposition;europe;signal processing;sparse matrices;nonnegative matrix factorization;missing data;random projections;low-rank matrix completion;blind source separation;big data"],"search_terms":["apply","random","projections","nonnegative","matrix","factorization","missing","entries","yahaya","puigt","delmaire","roussel"],"title":"How to Apply Random Projections to Nonnegative Matrix Factorization with Missing Entries?","year":2019,"dataSources":["NqWTiMfRR56v86wRs","r6oz3cMyC99QfiuHW"]}