Up: 2:1

Nonlinear support vector machines can systematically identify stocks with high and low future returns

Ramon Huerta; Fernando Corbacho; Charles Elkan

Algorithmic Finance (2013), 2:1, 45-58
DOI: 10.3233/AF-13016

Published: Abstract, PDF.
Archived: SSRN.

Abstract

This paper investigates the profitability of a trading strategy based on training a model to identify stocks with high or low predicted returns. A tail set is defined to be a group of stocks whose volatility-adjusted price change is in the highest or lowest quantile, for example the highest or lowest 5%. Each stock is represented by a set of technical and fundamental features computed using CRSP and Compustat data. A classifier is trained on historical tail sets and tested on future data. The classifier is chosen to be a nonlinear support vector machine (SVM) due to its simplicity and effectiveness. The SVM is trained once per month, in order to adjust to changing market conditions. Portfolios are formed by ranking stocks using the classifier output. The highest ranked stocks are used for long positions and the lowest ranked ones for short sales. The Global Industry Classification Standard is used to build a model for each sector such that a total of 8 long-short portfolios for Energy, Materials, Industrials, Consumer Discretionary, Consumer Staples, Health Care, Financials, and Information Technology are formed. The data range from 1981 to 2010. Without measuring trading costs, but using 91 day holding periods to minimize these, the strategy leads to annual excess returns (Jensen alpha) of 15% with volatilities under 8% using the top 25% of the stocks of the distribution for training long positions and the bottom 25% for the short ones.

Managing Editor

Philip Maymin

University of Bridgeport

Deputy Managing Editor

Jayaram Muthuswamy

Kent State University

Advisory Board

Kenneth J. Arrow

Stanford University

Herman Chernoff

Harvard University

David S. Johnson

AT&T Labs Research

Leonid Levin

Boston University

Myron Scholes

Stanford University

Michael Sipser

Massachusetts Institute of Technology

Richard Thaler

University of Chicago

Stephen Wolfram

Wolfram Research

Editorial Board

Associate Editors

Peter Bossaerts

California Institute of Technology

Emanuel Derman

Columbia University

Ming-Yang Kao

Northwestern University

Pete Kyle

University of Maryland

David Leinweber

Lawrence Berkeley National Laboratory

Richard J. Lipton

Georgia Tech

Avi Silberschatz

Yale University

Robert Webb

University of Virginia

Affiliate Editors

Giovanni Barone-Adesi

University of Lugano

Bruce Lehmann

University of California, San Diego

Unique Features of the Journal

Open access
Online articles are freely available to all.
No submission fees
There is no cost to submit articles for review. There will also be no publication or author fee for at least the first two volumes.
Authors retain copyright
Authors may repost their versions of the papers on preprint archives, or anywhere else, at any time.
Enhanced content
Enhanced, interactive, computable content will accompany papers whenever possible. Possibilities include code, datasets, videos, and live calculations.
Comments
Algorithmic Finance is the first journal in the Financial Economics Network of SSRN to allow comments.
Archives
The journal is published by IOS Press. In addition, the journal maintains an archive on SSRN.com.
Legal
While the journal does reserve the right to change these features at any time without notice, the intent will always be to provide the world's most freely and quickly available research on algorithmic finance.
ISSN
Online ISSN: 2157-6203
Print ISSN: 2158-5571