A three-state opinion formation model for financial markets

We propose a three-state microscopic opinion formation model for the purpose of simulating dynamics financial markets. In order to mimic heterogeneous composition mass investors in market, agent-based considers two different types traders: noise traders and contrarians. Agents are represented as nodes network interactions they can assume any three distinct possible states. The time evolution state an agent is dictated by probabilistic that include both local global influences. A trader subject interactions, tending majority its nearest neighbors, whilst contrarian interaction with behavior market whole, minority market. exhibits typical qualitative quantitative features real series, including distributions returns heavy tails, volatility clustering long-time memory absolute values returns. fitted means coupled Gaussian distributions, quantitatively revealing transitions between leptokurtic, mesokurtic platykurtic regimes terms non-linear statistical coupling which describes complexity system.