J.-P. Onnela,A. Chakraborti,K. Kaski,J. Kertesz,A. Kanto
Abstract:This paper introduces a new methodology for constructing a network of companies called a dynamic asset graph. This is similar to the dynamic asset tree studied recently, as both are based on correlations between asset returns. However, the new modified methodology does not, in general, lead to a tree but a graph, or several graphs that need not be inter-connected. The asset tree, due to the minimum spanning tree criterion, is forced to ``accept'' edge lengths that are far less optimal (longer) than the asset graph, thus resulting in higher overall length for the tree. The same criterion also causes asset trees to be more fragile in structure when measured by the single-step survival ratio. Over longer time periods, in the beginning the asset graph decays more slowly than the asset tree, but in the long-run the situation is reversed. The vertex degree distributions indicate that the possible scale free behavior of the asset graph is not as evident as it is in the case of the asset tree.
What problem does this paper attempt to address?
Judging from the provided text, this paper mainly focuses on the study of complex system behaviors in statistical mechanics. Specifically, the problems that this paper attempts to solve can be summarized as follows:
1. **Analysis of the Dynamic Characteristics of Complex Systems**:
- Researchers analyze the dynamic behaviors of complex systems by constructing time - series data. These time - series data include the changes in the interaction intensity between multiple nodes over time.
- The formula \( \rho_{ij}^t=\langle r_i^t r_j^t \rangle-\langle r_i^t \rangle \langle r_j^t \rangle \sqrt{[\langle (r_i^t)^2 \rangle-\langle r_i^t \rangle^2][\langle (r_j^t)^2 \rangle-\langle r_j^t \rangle^2]} \) is used to calculate the correlation between node \( i \) and node \( j \) at time \( t \).
- The distance \( d_{ij}^t = \sqrt{2(1 - \rho_{ij}^t)} \) is used to quantify the similarity or difference between nodes.
2. **Evolution of Network Structures**:
- The paper explores how to describe the structural changes of complex networks through time evolution. Each time point \( t \) corresponds to a network \( G_t=(V_t, E_t) \), where \( V_t \) represents the set of nodes and \( E_t \) represents the set of edges.
- By analyzing the network structures at different time points, researchers attempt to reveal the dynamic evolution laws of the system over time.
3. **Extraction of Statistical Features**:
- The paper also focuses on how to extract statistical features from time - series data, such as the distance distribution between nodes, the connectivity of the network, etc.
- These statistical features are helpful for understanding the internal mechanisms and behavior patterns of complex systems.
4. **Application Background**:
- Although the specific background is not elaborated in detail, this type of research is usually applied in multiple fields such as physics, biology, and sociology, and is of great significance especially in the study of multi - body systems, ecosystems, social networks, etc.
In summary, the main objective of this paper is to reveal the dynamic behaviors and internal mechanisms of complex systems through time - series data analysis and network structure evolution research. This not only contributes to theoretical research but also provides an important reference for practical applications.