Online Time Series Prediction with Missing Data

Oren Anava, Elad Hazan, Assaf Zeevi
Proceedings of the 32nd International Conference on Machine Learning, PMLR 37:2191-2199, 2015.

Abstract

We consider the problem of time series prediction in the presence of missing data. We cast the problem as an online learning problem in which the goal of the learner is to minimize prediction error. We then devise an efficient algorithm for the problem, which is based on autoregressive model, and does not assume any structure on the missing data nor on the mechanism that generates the time series. We show that our algorithm’s performance asymptotically approaches the performance of the best AR predictor in hindsight, and corroborate the theoretic results with an empirical study on synthetic and real-world data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v37-anava15, title = {Online Time Series Prediction with Missing Data}, author = {Anava, Oren and Hazan, Elad and Zeevi, Assaf}, booktitle = {Proceedings of the 32nd International Conference on Machine Learning}, pages = {2191--2199}, year = {2015}, editor = {Bach, Francis and Blei, David}, volume = {37}, series = {Proceedings of Machine Learning Research}, address = {Lille, France}, month = {07--09 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v37/anava15.pdf}, url = {https://proceedings.mlr.press/v37/anava15.html}, abstract = {We consider the problem of time series prediction in the presence of missing data. We cast the problem as an online learning problem in which the goal of the learner is to minimize prediction error. We then devise an efficient algorithm for the problem, which is based on autoregressive model, and does not assume any structure on the missing data nor on the mechanism that generates the time series. We show that our algorithm’s performance asymptotically approaches the performance of the best AR predictor in hindsight, and corroborate the theoretic results with an empirical study on synthetic and real-world data.} }
Endnote
%0 Conference Paper %T Online Time Series Prediction with Missing Data %A Oren Anava %A Elad Hazan %A Assaf Zeevi %B Proceedings of the 32nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2015 %E Francis Bach %E David Blei %F pmlr-v37-anava15 %I PMLR %P 2191--2199 %U https://proceedings.mlr.press/v37/anava15.html %V 37 %X We consider the problem of time series prediction in the presence of missing data. We cast the problem as an online learning problem in which the goal of the learner is to minimize prediction error. We then devise an efficient algorithm for the problem, which is based on autoregressive model, and does not assume any structure on the missing data nor on the mechanism that generates the time series. We show that our algorithm’s performance asymptotically approaches the performance of the best AR predictor in hindsight, and corroborate the theoretic results with an empirical study on synthetic and real-world data.
RIS
TY - CPAPER TI - Online Time Series Prediction with Missing Data AU - Oren Anava AU - Elad Hazan AU - Assaf Zeevi BT - Proceedings of the 32nd International Conference on Machine Learning DA - 2015/06/01 ED - Francis Bach ED - David Blei ID - pmlr-v37-anava15 PB - PMLR DP - Proceedings of Machine Learning Research VL - 37 SP - 2191 EP - 2199 L1 - http://proceedings.mlr.press/v37/anava15.pdf UR - https://proceedings.mlr.press/v37/anava15.html AB - We consider the problem of time series prediction in the presence of missing data. We cast the problem as an online learning problem in which the goal of the learner is to minimize prediction error. We then devise an efficient algorithm for the problem, which is based on autoregressive model, and does not assume any structure on the missing data nor on the mechanism that generates the time series. We show that our algorithm’s performance asymptotically approaches the performance of the best AR predictor in hindsight, and corroborate the theoretic results with an empirical study on synthetic and real-world data. ER -
APA
Anava, O., Hazan, E. & Zeevi, A.. (2015). Online Time Series Prediction with Missing Data. Proceedings of the 32nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 37:2191-2199 Available from https://proceedings.mlr.press/v37/anava15.html.

Related Material