Online Learning for Time Series Prediction

Abstract:

Online learning is a rich and fast-growing literature with algorithms
benefitting from regret guarantees that are often tight. Can we
leverage the online learning theory and algorithms to devise accurate
solutions for time series prediction in the stochastic setting?

This talk presents a series of theoretical and algorithmic solutions
addressing this question. It further shows how some notoriously
difficult time series problems such as model selection and ensemble
learning can be tackled using these ideas.

(joint work with Vitaly Kuznetsov)