Estimating Behavioral Agent-Based Models for Financial Markets through Machine Learning Surrogates

Document Type : Original Article

Author

1 Department of Business Administration, Faculty of Business Administration, Economics and Political Science, The British University in Egypt, ElSherouk City, Egypt. 2 Department of Socio-Computing, Faculty of Economics and Political Science, Cairo University, Cairo, Egypt.

Abstract

Traditional economic assumptions such as rational, 
representative agents and efficient market hypothesis failed to 
explain the macro-behavior of financial markets. On the other 
hand, agent-based approach proves high potentials in modeling 
bounded rational and heterogeneous micro-behaviors. This 
approach captures important stylized facts of financial markets. 
However, the high complexity of estimating agent-based models 
parameters precludes using these models in the forecasting 
process. This problem limits the applicability of agent-based 
models in decision making and policy formulation processes. 
Thereafter, this research aims at introducing a prospect for 
estimating agent-based models for financial markets through
surrogate modeling approach. Surrogate models are considered as 
novel parameter estimation method in economics though it is a 
well-defined method in engineering. Few efforts have been spent 
to estimate parameters using surrogate models.

Keywords

Files

Share

How to cite

Statistics