¶ Description
Community detection.
¶ Parameters
¶ Computations tab
Parameters:
- Output
- Community Detection (Newman's modularity)
- Algorithm
- Resolution parameter 'g'
- Number of hierarchical level to output
- Allow one community equal to complete network
- Treat disconnected components separately
- Social Leader Ranking (Sum-Over-All-Paths-Betweeness Centrality)
- parameters.CommunityDetection.computeCentrality
- Number of hirarchical level to output
¶ Connections tab
Parameters:
- Connection (Arcs) from A to B
- Column with ID A
- Detect A community
- Column with ID B
- Detect B community
- Weight of a connection A<->B
- Number of hierarchical level to output
- Allow self-loops arcs
¶ Advanced tab
Parameters:
- Iteration control at each hierarchical level
- Maximum number of iterations
- Iterate while
- Modularity increase of, at least
- there are at least % of nodes that changed from community
- Random seed
- nCpu for centrality computation
¶ About
The CommunityDetection action identifies clusters (communities) in a graph using Newman's modularity. It processes node connections to assign each node to one or more hierarchical community levels, supporting both standard modularity and social leader-based approaches.
¶ Input Example
The action expects graph edge data as input. You can use InlineTable or a .csv file with the following structure:
| Source | Target | Weight |
|---|---|---|
| 1 | 2 | 1 |
| 2 | 3 | 2 |
| 3 | 4 | 1 |
| 4 | 5 | 3 |
| 5 | 1 | 2 |
¶ Output Example
After processing, the output table shows each node's community membership at different levels:
| id | communityID_L1 | communityID_L2 |
|---|---|---|
| 1 | 0 | 0 |
| 2 | 1 | 1 |
| 3 | 1 | 1 |
| 4 | 2 | 2 |
| 5 | 2 | 2 |
¶ Use Cases
- Detecting communities in social networks
- Analyzing clustering in communication or transport networks
- Identifying modular structure in customer-product interactions
- Enhancing feature engineering for graph-based machine learning
¶ Notes
- All columns must be numeric (
Source,Target,Weight). - If the input includes disconnected components, enable the "Treat disconnected components separately" option.
- To avoid errors, ensure the column names are exactly mapped in the Connections tab.
¶ Screenshots
📸 Connections Configuration:
📸 Computation Configuration:
📸 Advanced Settings:
📸 Execution Result Table:
