Spectral Measures of Bipartivity in Complex Networks

Ernesto Estrada,Juan A. Rodriguez-Velazquez

DOI: https://doi.org/10.1103/PhysRevE.72.046105

2005-04-27

Abstract:We introduce a quantitative measure of network bipartivity as a proportion of even to total number of closed walks in the network. Spectral graph theory is used to quantify how close to bipartite a network is and the extent to which individual nodes and edges contribute to the global network bipartivity. It is shown that the bipartivity characterizes the network structure and can be related to the efficiency of semantic or communication networks, trophic interactions in food webs, construction principles in metabolic networks, or communities in social networks.

Statistical Mechanics,Physics and Society

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to quantify bipartivity in complex networks. Specifically, the author introduced a bipartivity measurement method based on the ratio of the number of even - length closed paths to the total number of closed paths in the network. By using spectral graph theory, the author not only quantified the overall bipartivity degree of the network, but also analyzed the contributions of individual nodes and edges to the bipartivity of the entire network. This method can be used to describe network structural characteristics and has potential applications in areas such as semantic or communication networks, food webs, metabolic networks, and social networks.

### Main contributions of the paper:

1. **Proposed a new bipartivity measurement method**: This method is based on the ratio of the number of even - length closed walks to the total number of closed paths in the network and can effectively quantify the bipartivity of the network.
2. **Application of spectral graph theory**: Using spectral graph theory, the bipartivity measurement is calculated through the eigenvalues of the adjacency matrix of the network, which makes the calculation process more straightforward.
3. **Local bipartivity measurement**: It not only provides the overall bipartivity measurement of the network, but also can calculate the contribution of each node and edge to the network bipartivity, which is very useful for understanding the details of network structure.
4. **Practical applications**: By analyzing different types of real - world networks, the application value of this measurement method in different fields, such as semantic networks, social networks, biological networks, and technological networks, is demonstrated.

### Formula summary:

- **Network bipartivity measurement**:
\[
\beta(G)=\frac{\sum_{j = 1}^N\cosh(\lambda_j)}{\sum_{j = 1}^N\cosh(\lambda_j)+\sum_{j = 1}^N\sinh(\lambda_j)}
\]
where \(\lambda_j\) is the eigenvalue of the network adjacency matrix.

- **Contribution of node \(i\) to network bipartivity**:
\[
\beta(i)=\frac{\sum_{j = 1}^N\left(\mathbf{v}_j(i)\right)^2\cosh(\lambda_j)}{\sum_{j = 1}^N\cosh(\lambda_j)}
\]
where \(\mathbf{v}_j(i)\) is the \(i\)-th component of the eigenvector \(\mathbf{v}_j\).

- **Contribution of edge \(e\) to network bipartivity**:
\[
\beta(e)=\beta(G)-\beta(G - e)
\]

These formulas and methods provide a powerful tool for understanding and analyzing the bipartivity of complex networks.