Exploring Graph Theory to Identify Network Similarities

[Jubeyer]

Hi,
I am very new in Graph Theory, and am currently trying to figure out it's potential to solve my research problem. Here it goes:

I have a large physical network (10000 nodes, around 14,000 edges), it can be represented by an undirected weighted graph. Many portions of the network might have some similar structure. Is there any way to identify the similar portions?

I am curious about any index which may tell something about topological similarity or similarity in weights, etc.

Regards,
Jubeyer

 

[Future Bruno]

One possible approach to this problem is the use of graph clustering algorithms. These algorithms attempt to group similar nodes together by analyzing the structure of the graph and the weights of the edges. Depending on the specific algorithm that you use, it may be able to identify portions of the graph that have similar topology or similar edge weights. You can find many different graph clustering algorithms available online, such as k-means clustering, hierarchical clustering, and spectral clustering. Additionally, many of these algorithms come with software packages that you can download and use to analyze your data.