Nonparametric sequential prediction of time series

TL;DR

This paper introduces nonparametric sequential prediction strategies for time series that combine multiple experts, demonstrating their universal consistency and superior performance over traditional ARMA models across various real-world datasets.

Contribution

The paper develops and analyzes nonparametric prediction methods based on expert combination, proving their universal consistency and showing they outperform ARMA models in practice.

Findings

Nonparametric strategies are more flexible and faster than ARMA.

They achieve lower normalized cumulative prediction error.

Strategies are universally consistent under minimal conditions.

Abstract

Time series prediction covers a vast field of every-day statistical applications in medical, environmental and economic domains. In this paper we develop nonparametric prediction strategies based on the combination of a set of 'experts' and show the universal consistency of these strategies under a minimum of conditions. We perform an in-depth analysis of real-world data sets and show that these nonparametric strategies are more flexible, faster and generally outperform ARMA methods in terms of normalized cumulative prediction error.