The Hierarchical Spectral Merger algorithm: A New Time Series Clustering Procedure

TL;DR

The paper introduces the Hierarchical Spectral Merger (HSM) algorithm for time series clustering, leveraging spectral theory to identify series with similar oscillations, demonstrated on oceanography and EEG data.

Contribution

The paper proposes a novel spectral-based hierarchical clustering method for time series, with an R implementation and applications to oceanographic and EEG datasets.

Findings

Effective clustering of time series with similar spectral features

Successful application to oceanography and EEG data

Provides a new spectral density estimation approach during clustering

Abstract

We present a new method for time series clustering which we call the Hierarchical Spectral Merger (HSM) method. This procedure is based on the spectral theory of time series and identifies series that share similar oscillations or waveforms. The extent of similarity between a pair of time series is measured using the total variation distance between their estimated spectral densities. At each step of the algorithm, every time two clusters merge, a new spectral density is estimated using the whole information present in both clusters, which is representative of all the series in the new cluster. The method is implemented in an R package HSMClust. We present two applications of the HSM method, one to data coming from wave-height measurements in oceanography and the other to electroencefalogram (EEG) data.