| Abstract [eng] |
Extreme, quake-like, events are far more common to financial and social systems than they should be according to the efficient market hypothesis and equilibrium system evolution models. These events, observed in social behavior of humans and animals, may be seen as directly related to the concepts of peer pressure and herding behavior. Modeling of the financial markets is the main objective of this dissertation, but the results are obtained in broader context of socio-economic modeling in general. The presented research is based on ideas and methods common to statistical physics are used in unison with contemporary socio-economic modeling frameworks, such as agent-based modeling and network theory. In this dissertation a series of models, growing in complexity, are proposed to better reproduce the empirical statistical features. The proposed consentaneous three state model, which is the main result of the dissertation, reproduces empirical statistical distribution and spectral density of absolute return. |
