Nonnegative Matrix Factorization for Time Series Recovery From a Few Temporal Aggregates

Jiali Mei, Yohann De Castro, Yannig Goude, Georges Hébrail
Proceedings of the 34th International Conference on Machine Learning, PMLR 70:2382-2390, 2017.

Abstract

Motivated by electricity consumption reconstitution, we propose a new matrix recovery method using nonnegative matrix factorization (NMF). The task tackled here is to reconstitute electricity consumption time series at a fine temporal scale from measures that are temporal aggregates of individual consumption. Contrary to existing NMF algorithms, the proposed method uses temporal aggregates as input data, instead of matrix entries. Furthermore, the proposed method is extended to take into account individual autocorrelation to provide better estimation, using a recent convex relaxation of quadratically constrained quadratic programs. Extensive experiments on synthetic and real-world electricity consumption datasets illustrate the effectiveness of the proposed method.

Cite this Paper

BibTeX
@InProceedings{pmlr-v70-mei17a, title = {Nonnegative Matrix Factorization for Time Series Recovery From a Few Temporal Aggregates}, author = {Jiali Mei and De Castro, Yohann and Yannig Goude and Georges H{\'e}brail}, booktitle = {Proceedings of the 34th International Conference on Machine Learning}, pages = {2382--2390}, year = {2017}, editor = {Precup, Doina and Teh, Yee Whye}, volume = {70}, series = {Proceedings of Machine Learning Research}, month = {06--11 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v70/mei17a/mei17a.pdf}, url = {https://proceedings.mlr.press/v70/mei17a.html}, abstract = {Motivated by electricity consumption reconstitution, we propose a new matrix recovery method using nonnegative matrix factorization (NMF). The task tackled here is to reconstitute electricity consumption time series at a fine temporal scale from measures that are temporal aggregates of individual consumption. Contrary to existing NMF algorithms, the proposed method uses temporal aggregates as input data, instead of matrix entries. Furthermore, the proposed method is extended to take into account individual autocorrelation to provide better estimation, using a recent convex relaxation of quadratically constrained quadratic programs. Extensive experiments on synthetic and real-world electricity consumption datasets illustrate the effectiveness of the proposed method.} }
Endnote
%0 Conference Paper %T Nonnegative Matrix Factorization for Time Series Recovery From a Few Temporal Aggregates %A Jiali Mei %A Yohann De Castro %A Yannig Goude %A Georges Hébrail %B Proceedings of the 34th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2017 %E Doina Precup %E Yee Whye Teh %F pmlr-v70-mei17a %I PMLR %P 2382--2390 %U https://proceedings.mlr.press/v70/mei17a.html %V 70 %X Motivated by electricity consumption reconstitution, we propose a new matrix recovery method using nonnegative matrix factorization (NMF). The task tackled here is to reconstitute electricity consumption time series at a fine temporal scale from measures that are temporal aggregates of individual consumption. Contrary to existing NMF algorithms, the proposed method uses temporal aggregates as input data, instead of matrix entries. Furthermore, the proposed method is extended to take into account individual autocorrelation to provide better estimation, using a recent convex relaxation of quadratically constrained quadratic programs. Extensive experiments on synthetic and real-world electricity consumption datasets illustrate the effectiveness of the proposed method.
APA
Mei, J., De Castro, Y., Goude, Y. & Hébrail, G.. (2017). Nonnegative Matrix Factorization for Time Series Recovery From a Few Temporal Aggregates. Proceedings of the 34th International Conference on Machine Learning, in Proceedings of Machine Learning Research 70:2382-2390 Available from https://proceedings.mlr.press/v70/mei17a.html.

Related Material