Clustering Financial Time Series: How Long is Enough?
Machine Learning
2016-04-18 v2 Statistical Finance
Abstract
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.
Cite
@article{arxiv.1603.04017,
title = {Clustering Financial Time Series: How Long is Enough?},
author = {Gautier Marti and Sébastien Andler and Frank Nielsen and Philippe Donnat},
journal= {arXiv preprint arXiv:1603.04017},
year = {2016}
}
Comments
Accepted at IJCAI 2016