Online Learning for Time Series Prediction

We address the problem of predicting a time series using the ARMA (autoregressive moving average) model, under minimal assumptions on the noise terms. Using regret minimization techniques, we develop eective online learning algorithms for the prediction problem, without assuming that the noise terms are Gaussian, identically distributed or even independent. Furthermore, we show that our algorithm’s performances asymptotically approaches the performance of the best ARMA model in hindsight.

[1]  P. Young,et al.  Time series analysis, forecasting and control , 1972, IEEE Transactions on Automatic Control.

[2]  R. Engle Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation , 1982 .

[3]  Jonathan D. Cryer,et al.  Time Series Analysis , 1986, Encyclopedia of Big Data.

[4]  A. El-Shaarawi,et al.  ARMA Models with Double-exponentially Distributed Noise , 1989 .

[5]  Richard A. Davis,et al.  Time Series: Theory and Methods , 2013 .

[6]  David J. Thomson,et al.  Jackknifing multiple-window spectra , 1994, Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing.

[7]  M. Tiku,et al.  Time Series Models in Non‐Normal Situations: Symmetric Innovations , 2000 .

[8]  Martin Zinkevich,et al.  Online Convex Programming and Generalized Infinitesimal Gradient Ascent , 2003, ICML.

[9]  Gábor Lugosi,et al.  Prediction, learning, and games , 2006 .

[10]  Feng Ding,et al.  Performance analysis of estimation algorithms of nonstationary ARMA processes , 2006, IEEE Transactions on Signal Processing.

[11]  竹安 数博,et al.  Time series analysis and its applications , 2007 .

[12]  Elad Hazan,et al.  Logarithmic regret algorithms for online convex optimization , 2006, Machine Learning.

[13]  Seshadhri Comandur,et al.  Efficient learning algorithms for changing environments , 2009, ICML '09.

[14]  Elad Hazan,et al.  learning algorithms for changing environments , 2009 .