Abstract | ||
---|---|---|
We present data-dependent learning bounds for the general scenario of non-stationary non-mixing stochastic processes. Our learning guarantees are expressed in terms of a data-dependent measure of sequential complexity and a discrepancy measure that can be estimated from data under some mild assumptions. We use our learning bounds to devise new algorithms for non-stationary time series forecasting for which we report some preliminary experimental results. |
Year | Venue | Field |
---|---|---|
2015 | Annual Conference on Neural Information Processing Systems | Time series,Mathematical optimization,Algorithmic learning theory,Learning theory,Computer science,Empirical risk minimization,Stochastic process,Algorithm,Generalization error,Artificial intelligence,Computational learning theory,Machine learning |
DocType | Volume | ISSN |
Conference | 28 | 1049-5258 |
Citations | PageRank | References |
6 | 0.59 | 12 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Vitaly Kuznetsov | 1 | 68 | 9.33 |
Mehryar Mohri | 2 | 4502 | 448.21 |