Reproduce stylized facts of artificial financial market and comparison with real data

Abstract

Agent-based computational models represent a big challenge in many disciplines. A vital approach receiving much interest is agent-based models, which gives a new area providing some ways to tackle some of the restrictions of the analytical models in finance. The aim of our research is to contribute to the behavioral finance and agent-based artificial markets by studying their market-wise implications using computational simulations. We investigate and analyze the behavioral foundations of the stylized facts of empirical data such as that characterize real data in financial markets. Our results confirm the existence of most the stylized facts such as leptokurtosis, non-independently distributed, and volatility clustering. From this attention, the artificial financial market will for all time be evaluated in order to have explication about market dynamics in Tunisian financial market.