Global decomposition of networks into multiple cores formed by local hubs

Wonhee Jeong,Unjong Yu,Sang Hoon Lee

2024-09-20

Abstract:Networks are ubiquitous in various fields, representing systems where nodes and their interconnections constitute their intricate structures. We introduce a network decomposition scheme to reveal multiscale core-periphery structures lurking inside, using the concept of locally defined nodal hub centrality and edge-pruning techniques built upon it. We demonstrate that the hub-centrality-based edge pruning reveals a series of breaking points in network decomposition, which effectively separates a network into its backbone and shell structures. Our local-edge decomposition method iteratively identifies and removes locally least important nodes, and uncovers an onion-like hierarchical structure as a result. Compared with the conventional $k$-core decomposition method, our method based on relative information residing in local structures exhibits a clear advantage in terms of discovering locally crucial substructures. Furthermore, we introduce the core-periphery score to properly separate the core and periphery with our decomposition scheme. By extending the method combined with the network community structure, we successfully detect multiple core-periphery structures by decomposition inside each community. Moreover, the application of our decomposition to supernode networks defined from the communities reveals the intricate relation between the two representative mesoscale structures.

Physics and Society,Statistical Mechanics,Social and Information Networks

What problem does this paper attempt to address?

### Problems the paper attempts to solve

The paper aims to solve the core - periphery structure identification problem in network structure analysis. Specifically, the paper proposes a new network decomposition scheme, which reveals the hidden multi - scale core - periphery structures in the network through the concept of hub centrality and the edge - pruning technique based on it. Different from the traditional k - core decomposition method based on global degree, this method can more effectively discover local key sub - structures and provide more detailed network organization information.

### Main contributions

1. **Introduction of local node centrality**: The paper proposes a network decomposition method based on local node centrality. This method can identify local hub nodes in the network and use these nodes to gradually decompose the network.

2. **Edge - pruning technique**: By calculating the importance of the edge connecting two nodes (based on the product of the local node centralities of the nodes), the paper proposes an edge - pruning technique that can effectively separate the core part and the periphery part of the network.

3. **Multi - level core - periphery structure**: The paper shows how to reveal the multi - level core - periphery structure of the network by iteratively applying the above - mentioned decomposition method, forming an onion - like hierarchical structure.

4. **Core - periphery score**: In order to better separate the core and the periphery, the paper introduces a core - periphery score and detects multiple core - periphery structures within each community.

5. **Application of super - node network**: The paper also explores the application of the decomposition method to the super - node network defined by communities, revealing the complex relationship between two representative meso - scale structures.

### Method overview

1. **Definition of local node centrality**: Local node centrality is defined as the normalized value of the degree ranking of a node in the node group composed of itself and its nearest neighbor nodes. This concept can identify locally important nodes.

2. **Edge - pruning process**: Determine the importance of each edge by calculating the product of the local node centralities of the edge (\( E_{ij}=\psi_i\psi_j \)). Then, prune the edges in order from low to high importance until the network is decomposed into the core part and the periphery part.

3. **Natural break point**: During the pruning process, the paper discovers a natural break point (cusp point), that is, when the proportion of removed edges reaches a certain value, the size of the giant component of the network suddenly decreases. This break point is used to determine the core part of the network.

4. **Iterative decomposition**: By repeatedly applying the above process, the multi - level core parts of the network can be gradually extracted, forming an onion - like hierarchical structure.

### Application examples

The paper demonstrates the effectiveness of this method through several examples of actual networks (such as the cooperation network in the field of computational geometry, the email network, the Brightkite social network, and the Wikipedia word network). The results show that this method can more clearly reveal the multi - level core - periphery structure in the network and provide more abundant network organization information compared with the traditional k - core decomposition method.

### Conclusion

By introducing local node centrality and the edge - pruning technique based on it, the paper proposes a new network decomposition method that can more effectively identify the multi - level core - periphery structures in the network. This method is not only of great theoretical significance but also provides new tools and perspectives for practical applications.