Artículos

Análisis de Mercados Financieros utilizando Técnicas de Inteligencia Artificial

Analysis of Financial Markets using Artificial Intelligence Techniques

José Luis Gordillo¹ y Christopher R. Stephens²

1 Dirección General de Servicios de Cómputo Académico, UNAM
Circuito Exterior s/n, Ciudad Universitaria CP 04510
2 Instituto de Ciencias Nucleares, UNAM
Circuito Exterior s/n, Ciudad Universitaria
Apartado postal 70–543, CP 04510
México, DF, México

E–mail: jlgr@super.unam.mx, stephens@nuclecu.unam.mx

Artículo recibido en Junio 03, 2002; aceptado en Abril 15, 2003

Resumen

Las técnicas actuales de Inteligencia Artificial (IA), en particular los algoritmos evolutivos y los agentes inteligentes, permiten modelar sistemas de muy diversos tipos, entre ellos mercados financieros. Un modelo computacional de un mercado financiero utiliza varios elementos de AI, tales como la computaci6n evolutiva y agentes artificiales. El resultado –un mercado financiero artificial (MFA)– sirve como un laboratorio útil para el estudio de los mercados reales. En este trabajo presentamos un nuevo MFA, construido con el objetivo de incluir elementos que han sido descartados en otros modelos, tales como diferentes modelos de organización del mercado y la presencia de participantes especiales, tales como los market makers, que son agentes especiales encargados de proveer liquidez al mercado. Los resultados presentados en este trabajo incluyen el efecto de diferentes modelos de organización, de diferentes usos de información, la eficiencia de diferentes métodos de adaptación y de la presencia de market makers. En este último caso, cabe enfatizar el hecho de tener un MFA cuya evolución es completamente autónoma.

Palabras clave: Inteligencia Artificial, Agentes, Mercados Financieros, Algoritmos Evolutivos, Market Markers.

Abstract

Present day techniques from artificial intelligence (AI), in particular evolutionary algorithms and intelligent agents, can be used to model many different types of system, including financial markets. A computational model of a financial market uses various AI elements, such as evolutionary computation and artificial agents. The result –an artificial financial market (AFM)– can serve as a useful "laboratory" for studying the behavior of real financial markets. In the present work we present a new AFM, designed with the objective of including elements that have been relatively neglected in other models, such as different organizational models for the market and the presence of priviliges participants such as market makers who have the task of providing market liquidity. The results we present in this article include: the effect of different organizational models, of different exploitation of information, the efficiency of different types of adaptation and the presence of market makers. In this latter case we emphasize that the resulting market has a completely autonomous evolution.

Keywords: Artificial Intelligence, Agents, Financial Markets, Evolutionary Algorithms, Market Markers.

DESCARGAR ARTICULO EN FORMATO PDF

Agradecimientos

Los autores agradecen el soporte de DGAPA, UNAM, proyecto IN100201; de igual manera agradecen el apoyo brindado por el Departamento de Supercómputo de la UNAM.

Referencias

Bedau M. y Joshi S., "An Explanation of Generic Behavior in an Evolving Financial Market", Santa Fe preprint 98–12–114E, Santa Fe Institute, 1998.        [ Links ]

Bradshaw J., "An introduction to software agents", AAAI Press. 1997.        [ Links ]

Noriega P., "Agent Mediated: The Fishmarket Metaphor", Ph. D. Thesis, Universitat Autònoma de Barcelona. 1997.        [ Links ]

Wooldridge M., "Reasoning about Rational Agents", The MIT Press, Cambridge Massachussets, London, England. 2000.        [ Links ]

Wooldridge M., Jennings N. R. y Kinny D., "The Gaia Methodology for Agent–Oriented Analysis and Design", Journal of Autonomous Agents and MultyAgent Systems, 3(3): 285–312, 2000.        [ Links ]

Campbell J.Y., Lo A.W. y MacKinley A.C., "The Econometrics of Financial Markets", Princeton University Press, EU, 1997.        [ Links ]

Chan N., Lebaron B., Lo A. y Poggio T., "Agent based models of financial markets: a comparison with experimental markets", Working Paper, Brandeis University, 1999.        [ Links ]

Farmer J. D., "Market force, ecology and evolution", Working Paper 98–12–117E, Santa Fe Institute, 1998.        [ Links ]

Goldberg D.G., "Genetic Algorithms in Search, Optimization and Machine Learning", Addison Wesley, 1989.        [ Links ]

Gopikrishnan P., Plerou V., Nunes L., Mayer M. y Stanley E., "Scaling of the distribution of fluctuations of financial market indices", Physical Review, Vol. 60. 1999.        [ Links ]

Gordillo J .L., "Análisis de Mercados Financieros mediante el Mercado Financiero Artificial NNCP", IIMAS–UNAM, 2000.        [ Links ]

Gordillo J.L y Stephens C.R., "Strategy Adaptation and the Role of Information in an Artificial Financial Market", publicado en Late Breaking Papers, GECCO 2001, San Francisco, Julio 2001.        [ Links ]

Krahnen J. y Weber M.. "Marketing in the laboratory: Does competition matter?", Finance Working Paper Series #4, Universität Frankfurt. 1998.        [ Links ]

Lebaron B., "Evolution and time horizons in an agent based stock market", Working Paper, Brandeis University, 1999.        [ Links ]

Lebaron B., "Building financial markets with artificial agents: desired goals and present techniques" , Computational Markets, MIT Press, 1999.        [ Links ]

Palmer R.G., Arthur W.B., Holland J.H., Lebaron B. y Tayler P., "Artificial economic life: a simple model of a stock market", Physica D 73, 1994.        [ Links ]

O'Hara M., "Market Microstructure Theory", Blackwell Publishers Inc, Malden MA, 1997.        [ Links ]

Stephens C.R., Waelbroeck H., Sudarsky D. y Zevtuche F., "Mercados Financieros Adaptivos", Soluciones Avanzadas 56, 1998.        [ Links ]

Páginas Electrónicas

Laboratory for Financial Engineering. Artificial Markets Project. MIT. http://lfe.mit.edu/research/artificial–mkts.html.        [ Links ]

Virtual Financial Markets. Center for Biological and Computational Learning. MIT. http://www.ai.mit.edu/projects/cbcl/res_area/finance/finance.html.        [ Links ]

Agent Based Computational Economics. Department of Economics. Iowa State University. http://www.econ.iastate.edu/tesfasti.        [ Links ]