Clustering Financial Time Series: How Long is Enough?

by   Gautier Marti, et al.

Researchers have used from 30 days to several years of daily returns as source data for clustering financial time series based on their correlations. This paper sets up a statistical framework to study the validity of such practices. We first show that clustering correlated random variables from their observed values is statistically consistent. Then, we also give a first empirical answer to the much debated question: How long should the time series be? If too short, the clusters found can be spurious; if too long, dynamics can be smoothed out.


page 2

page 5

page 6


Independence clustering (without a matrix)

The independence clustering problem is considered in the following formu...

Superstatistics with cut-off tails for financial time series

Financial time series have been investigated to follow fat-tailed distri...

Clustering Methods Assessment for Investment in Zero Emission Neighborhoods Energy System

This paper investigates the use of clustering in the context of designin...

Multifractality in time series is due to temporal correlations

Based on the rigorous mathematical arguments formulated within the Multi...

Toward a generic representation of random variables for machine learning

This paper presents a pre-processing and a distance which improve the pe...

Dynamic time series clustering via volatility change-points

This note outlines a method for clustering time series based on a statis...

Please sign up or login with your details

Forgot password? Click here to reset