Fast time-series prediction using high-dimensional data: Evaluating confidence interval credibility

Yoshito Hirata

Phys. Rev. E 89, 052916 – Published 29 May 2014

Abstract

I propose an index for evaluating the credibility of confidence intervals for future observables predicted from high-dimensional time-series data. The index evaluates the distance from the current state to the data manifold. I demonstrate the index with artificial datasets generated from the Lorenz'96 II model [Lorenz, in Proceedings of the Seminar on Predictability, Vol. 1 (ECMWF, Reading, UK, 1996), p. 1], the Lorenz'96 I model [Hansen and Smith, J. Atmos. Sci. 57, 2859 (2000)], and the coupled map lattice, and a real dataset for the solar irradiation around Japan.