Dynamic time series clustering via volatility change-points

TL;DR

This paper presents a dynamic clustering method for time series based on volatility change-points, using a statistical model that captures shifts in volatility and updates groupings as new data arrive.

Contribution

It introduces a novel online clustering approach leveraging change-point detection in volatility, integrating a probabilistic metric between posterior distributions.

Findings

Effective clustering of S&P 500 returns demonstrated

Method captures coincident volatility shifts in time series

Clustering updates dynamically with incoming data

Abstract

This note outlines a method for clustering time series based on a statistical model in which volatility shifts at unobserved change-points. The model accommodates some classical stylized features of returns and its relation to GARCH is discussed. Clustering is performed using a probability metric evaluated between posterior distributions of the most recent change-point associated with each series. This implies series are grouped together at a given time if there is evidence the most recent shifts in their respective volatilities were coincident or closely timed. The clustering method is dynamic, in that groupings may be updated in an online manner as data arrive. Numerical results are given analyzing daily returns of constituents of the S&P 500.