A New Method For Dynamic Stock Clustering Based On Spectral Analysis

A New Method For Dynamic Stock Clustering Based On Spectral Analysis In this paper, we propose a new method to classify the stock cluster based on the motions of stock returns. Specifically, there are three criteria: (1) The positive or negative signs of elements in the eigenvector of correlation matrix indicate the response direction of individual stocks. (2) The components are included based on the sequence of corresponding eigenvalue magnitudes from large to small. (3) All the elements in the cluster representing individual stocks should have same signs across the components included in the classification process. With the number of vectors included in the process increasing, a speed-up process for cluster number is identified. We interpret this phenomenon as a phase transition from a state dominated by meaningful information to one dominated by trivial information. And a critical point exists in this process. The sizes of clusters near this critical point can be regarded as a power law distribution. The critical exponent evolvement indicates structure of the market. The vector number at this point can be adopted to classify the stock clusters. We analyze the cross-correlation matrices of stock logarithm returns of both China and US stock market for the period from January 4, 2005 to December 31, 2010. The period includes the anomalies time of financial crisis. The number of clusters in financial and technology sectors is further examined to reveal the varying feather of traditional industries. Distinct patterns of industrial differentiation between China and US have been found according to our study. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Computational Economics Springer Journals

A New Method For Dynamic Stock Clustering Based On Spectral Analysis

Loading next page...
 
/lp/springer_journal/a-new-method-for-dynamic-stock-clustering-based-on-spectral-analysis-DfNzxxDLBg
Publisher
Springer US
Copyright
Copyright © 2016 by Springer Science+Business Media New York
Subject
Economics; Economic Theory/Quantitative Economics/Mathematical Methods; Computer Appl. in Social and Behavioral Sciences; Operations Research/Decision Theory; Behavioral/Experimental Economics; Math Applications in Computer Science
ISSN
0927-7099
eISSN
1572-9974
D.O.I.
10.1007/s10614-016-9589-9
Publisher site
See Article on Publisher Site

Abstract

In this paper, we propose a new method to classify the stock cluster based on the motions of stock returns. Specifically, there are three criteria: (1) The positive or negative signs of elements in the eigenvector of correlation matrix indicate the response direction of individual stocks. (2) The components are included based on the sequence of corresponding eigenvalue magnitudes from large to small. (3) All the elements in the cluster representing individual stocks should have same signs across the components included in the classification process. With the number of vectors included in the process increasing, a speed-up process for cluster number is identified. We interpret this phenomenon as a phase transition from a state dominated by meaningful information to one dominated by trivial information. And a critical point exists in this process. The sizes of clusters near this critical point can be regarded as a power law distribution. The critical exponent evolvement indicates structure of the market. The vector number at this point can be adopted to classify the stock clusters. We analyze the cross-correlation matrices of stock logarithm returns of both China and US stock market for the period from January 4, 2005 to December 31, 2010. The period includes the anomalies time of financial crisis. The number of clusters in financial and technology sectors is further examined to reveal the varying feather of traditional industries. Distinct patterns of industrial differentiation between China and US have been found according to our study.

Journal

Computational EconomicsSpringer Journals

Published: Jun 1, 2016

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off