Time series prediction and online learning

Vitaly Kuznetsov, Mehryar Mohri
29th Annual Conference on Learning Theory, PMLR 49:1190-1213, 2016.

Abstract

We present a series of theoretical and algorithmic results combining the benefits of the statistical learning approach to time series prediction with that of on-line learning. We prove new generalization guarantees for hypotheses derived from regret minimization algorithms in the general scenario where the data is generated by a non-stationary non-mixing stochastic process. Our theory enables us to derive model selection techniques with favorable theoretical guarantees in the scenario of time series, thereby solving a problem that is notoriously difficult in that scenario. It also helps us devise new ensemble methods with favorable theoretical guarantees for the task of forecasting non-stationary time series.

Cite this Paper

BibTeX
@InProceedings{pmlr-v49-kuznetsov16, title = {Time series prediction and online learning}, author = {Kuznetsov, Vitaly and Mohri, Mehryar}, booktitle = {29th Annual Conference on Learning Theory}, pages = {1190--1213}, year = {2016}, editor = {Feldman, Vitaly and Rakhlin, Alexander and Shamir, Ohad}, volume = {49}, series = {Proceedings of Machine Learning Research}, address = {Columbia University, New York, New York, USA}, month = {23--26 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v49/kuznetsov16.pdf}, url = {https://proceedings.mlr.press/v49/kuznetsov16.html}, abstract = {We present a series of theoretical and algorithmic results combining the benefits of the statistical learning approach to time series prediction with that of on-line learning. We prove new generalization guarantees for hypotheses derived from regret minimization algorithms in the general scenario where the data is generated by a non-stationary non-mixing stochastic process. Our theory enables us to derive model selection techniques with favorable theoretical guarantees in the scenario of time series, thereby solving a problem that is notoriously difficult in that scenario. It also helps us devise new ensemble methods with favorable theoretical guarantees for the task of forecasting non-stationary time series. } }
Endnote
%0 Conference Paper %T Time series prediction and online learning %A Vitaly Kuznetsov %A Mehryar Mohri %B 29th Annual Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2016 %E Vitaly Feldman %E Alexander Rakhlin %E Ohad Shamir %F pmlr-v49-kuznetsov16 %I PMLR %P 1190--1213 %U https://proceedings.mlr.press/v49/kuznetsov16.html %V 49 %X We present a series of theoretical and algorithmic results combining the benefits of the statistical learning approach to time series prediction with that of on-line learning. We prove new generalization guarantees for hypotheses derived from regret minimization algorithms in the general scenario where the data is generated by a non-stationary non-mixing stochastic process. Our theory enables us to derive model selection techniques with favorable theoretical guarantees in the scenario of time series, thereby solving a problem that is notoriously difficult in that scenario. It also helps us devise new ensemble methods with favorable theoretical guarantees for the task of forecasting non-stationary time series.
RIS
TY - CPAPER TI - Time series prediction and online learning AU - Vitaly Kuznetsov AU - Mehryar Mohri BT - 29th Annual Conference on Learning Theory DA - 2016/06/06 ED - Vitaly Feldman ED - Alexander Rakhlin ED - Ohad Shamir ID - pmlr-v49-kuznetsov16 PB - PMLR DP - Proceedings of Machine Learning Research VL - 49 SP - 1190 EP - 1213 L1 - http://proceedings.mlr.press/v49/kuznetsov16.pdf UR - https://proceedings.mlr.press/v49/kuznetsov16.html AB - We present a series of theoretical and algorithmic results combining the benefits of the statistical learning approach to time series prediction with that of on-line learning. We prove new generalization guarantees for hypotheses derived from regret minimization algorithms in the general scenario where the data is generated by a non-stationary non-mixing stochastic process. Our theory enables us to derive model selection techniques with favorable theoretical guarantees in the scenario of time series, thereby solving a problem that is notoriously difficult in that scenario. It also helps us devise new ensemble methods with favorable theoretical guarantees for the task of forecasting non-stationary time series. ER -
APA
Kuznetsov, V. & Mohri, M.. (2016). Time series prediction and online learning. 29th Annual Conference on Learning Theory, in Proceedings of Machine Learning Research 49:1190-1213 Available from https://proceedings.mlr.press/v49/kuznetsov16.html.

Related Material