A comparative study of several parametric and semiparametric approaches for time series classification

Abstract

Several non-parametric statistics originally designed to test the equality of the log-spectra of two stochastic processes are proposed as dissimilarity measures between two time series. Their behavior in time series clustering is analyzed throughout a simulation study, and compared with the performance of several model-free and model-based dissimilarity measures. Up to three different classification settings are considered: (1) to distinguish between stationary and non-stationary time series, (2) to classify different ARMA processes and (3) to classify several non-linear time series models. As it was expected, the performance of a particular dissimilarity metric strongly depended on the type of processes subjected to clustering. Among all the measures studied, the non-parametric distances showed the most robust behaviour. © Springer-Verlag Berlin Heidelberg 2010.