Clustering Evolutionary Stock Market Model

TL;DR

This paper introduces a dynamic clustering model of stock markets that captures fluctuations and large events, showing that price returns follow a Lévy distribution with exponential truncation, reflecting complex market behaviors.

Contribution

It presents a novel agent-based clustering model that simulates market fluctuations, including agent entry and exit, to better mimic real stock market dynamics.

Findings

Large market fluctuations are frequent in the model.

Stock price returns follow a Lévy distribution with exponential truncation.

The model reproduces key statistical features of real stock markets.

Abstract

As a typical representation of complex networks studied relatively thoroughly, financial market presents some special details, such as its nonconservation and opinions spreading. In this model, agents congregate to form some clusters, which may grow or collapse with the evolution of the system. To mimic an open market, we allow some ones participate in or exit the market suggesting that the number of the agents would fluctuate. Simulation results show that the large events are frequent in the fluctuations of the stock price generated by the artificial stock market when compared with a normal process and the price return distribution is a \emph{l\'{e}vy} distribution in the central part followed by an approximately exponential truncation.