Historical network mapping: clusters variations

CorText Manager Q&A forumCategory: Network mappingHistorical network mapping: clusters variations
Beadavi asked 3 months ago

I used the network mapping script to create a historical map from the cited references of a WOS database.
I get 5 months apart different representations and a different list of references while the parameters and the corpus used are the same.
Do you know where this can come from?
I thank you in advance

1 Answers
Lionel Staff answered 3 months ago

Dear Béatrice,
Not sure exactly what you mean: could it come from some overlapping period?
If it is something else, could you try to clarify your question a little?

Beadavi replied 3 months ago

Thank you for your answer. I did not define periods, I left the default parameters (louvain, automatic detection of the proximity measure).
I get a different number of clusters for the same analysis on the same corpus, and the list of references represented on the radar is not the same. In the parameters that can be consulted after the analysis, I get respectively :
After filtering, the network (period:1999_2019) has 15 nodes and 56 edges.
2 clusters
After filtering, the network (period:1999_2021) has 15 nodes and 64 edges.
3 clusters