@@ -87,3 +87,47 @@ The configuration of the nDR (n>1) method:
**depth**: specifies the portion of the base graph that is going to be re-clustered.
**clusterIdPrefix**: The prefix string for clusterIds of newly clustered entities.
### Incremental linking configurations
The linking configuration of the incremental matching tasks needs to be changed as follows:
When using MBM method:
```json
{
"keepCurrentEdges":"false",
"recomputeSimilarityForCurrentEdges":"false",
"tagNewLink":true,
"blockingComponents":[
...
],
"similarityComponents":[
...
],
"selectionComponent":{
...
}
}
```
**tagNewLink**: Specifies whether the edge between newly added entities (vertices) should be tagged with a property value 'newLink', regarding the edge status.
When using 1DR or nDR method:
```json
{
"keepCurrentEdges":"true",
"recomputeSimilarityForCurrentEdges":"true",
"tagNeighborCluster":true,
"blockingComponents":[
...
],
"similarityComponents":[
...
],
"selectionComponent":{
...
}
}
```
**tagNeighborCluster**: Specifies whether the neighbor cluster-id of a new entity (vertex) should be stored as a propertyValue in the connecting edge.
For nDR methods, we need to keep the previously computed edges for reclustering step.
