Hello everyone, I am studying a network composed of 123 nodes and 3211 edges. The network is disconnected, with connection weights ranging from 1 to 140 (the number of elements shared between nodes). In my network, the weights represent an affinity between nodes. My goal is to identify groups within the network. Since the modularity is very low (applying the Leiden algorithm with a resolution of 1, its value is 0.124), I intend to identify the nodes that most facilitate the flow of information within the network and eliminate them (removing outliers). I attempted to calculate the closeness betweenness using the inverses of the weights (1/matrix and replacing the Inf values with 0), but the result was that almost all nodes had a value of 0. Which centrality measure would be more appropriate in my case?
Related Topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
Documentation of the algorithm of eigen_centrality | 5 | 838 | 20 December 2021 | |
Weighted edges in centr_degree? | 2 | 697 | 27 May 2023 | |
Help seeking on calculating betweenness centrality with valued ties | 16 | 3102 | 11 June 2020 | |
Subgraph centrality for weighted graphs | 8 | 409 | 11 May 2022 | |
warning message in community detection with cluster_edge_betweenness algorithm | 8 | 1111 | 24 July 2023 |