2007/01/31 by D. Horvath, Denis Horváth, Z. Kuscsik +1 · 3 citations
Computer Science · Economics, Econometrics and Finance · Mathematics · Physics and Astronomy · #Artificial intelligence #Autoregressive conditional heteroskedasticity #Cluster analysis #Complex Systems and Time Series Analysis #Complex network #Computer science #Econometrics #Economics #Mathematics #Network dynamics #Nonlinear Dynamics and Pattern Formation #Opinion Dynamics and Social Influence #Random walk #Stock market #Stylized fact #Volatility (finance) #Volatility clustering #physics.soc-ph #q-fin.ST
paper · pdf · doi:10.1142/s0129183107011388
published in International Journal of Modern Physics C 18(08), 1361-1374 (World Scientific) · 13 pages, 5 figures, accepted in IJMPC, references added, minor changes in model, new results and modified figures
arxiv created 2007/03/05 · openalex publication_date 2007/08/01 · arxiv updated 2009/12/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/05
The agent-based model of stock price dynamics on a directed evolving complex network is suggested and studied by direct simulation. The stationary regime is maintained as a result of the balance between the extremal dynamics, adaptivity of strategic variables and reconnection rules. The inherent structure of node agent "brain" is modeled by a recursive neural network with local and global inputs and feedback connections. For specific parametric combination the complex network displays small-world phenomenon combined with scale-free behavior. The identification of a local leader (network hub, agent whose strategies are frequently adapted by its neighbors) is carried out by repeated random walk process through network. The simulations show empirically relevant dynamics of price returns and volatility clustering. The additional emerging aspects of stylized market statistics are Zipfian distributions of fitness.