Strange clusters
A widely used tool for visual exploration of scientific literature.
Brought to you by:
citespace
Menu
▾
▴
Dear Dr. Chen,
When I run a project I obtain some strange clusters. I do not know how to solve this or if this is important to follow with the analysis.
Image 01 shows a screen of CiteSpace Cluster Explorer. There you can see, as example, Cluster ID 8. This is normal.
But Image 02 shows a screen of two other clusters (Clusters ID 19 and 32) that are totally different as yo ucan see in the CITING ARTICLES part of the screen.
Could you be so kind to say me if this is normal? Or if this is not normal, where can I see to solve it?
Thanks
Miquel
Last edit: Miquel Angel Plaza 2021-03-17
When you use the function "Save Cluster Information", there is a dialog box asks you "Found an existing clusters folder. Is it OK to delete them?" You should select Yes. Otherwise, you may get a problem like the one you have.
To solve the problem, use the same function again and this time select "Yes".
Last edit: Chaomei Chen 2021-03-17
Thanks, Dr. Chen, but I followed your instructions and the result is the same. The cluster is still strange or at least, the info showed is strange
Go to your project folder, delete the folder called clusters, then try again.
It might also help if you shared some more information... Maybe from the narratives? And so what exactly do you find 'strange'?
Last edit: Stephan De Spiegeleire 2021-03-19
I do not know if it is really strange or not, and if this is a problem to follow with the analysis or not. I'm a beginner with CiteSpace and I'm trying to understand it.
If you see image 01.png in my first message you can see what I think are normal clusters when I run the option "Custer Explorer". In this image you can see the information for cluster 8. In "Citing articles" you can see numeric data in columns COVERAGE, GCS and LCS and bibliographic data in BIBLIOGRAPHY. column This happens also with clusters 4, 5, 6, 7, 16 and 17 in what I think are normal clusters.
But this is different with clusters 19 and 32. If you open image 02.png you can see totally different information. Column COVERAGE has information like terms... and columns GCS, LCS and BIBLIOGRAPHY are empty. For me these are strange clusters.
Is this normal? Can I follow with the analysis? Or must I review something and where?
I have done what you said, but I have still he same strange cluster
Hello. I would like to know if is it normal to obtain clusters like #19 and #32 of the following image. At TOP TERMS (LSI) column they appear as only three dots ...
And they have identical LLR terms, same info at COVERAGE section. And they do not have GCS, LCS not Bibliography info.
I have deleted Cluster subfolder in Projects folder... but I obtain the same info.
Yours
They look odd. Can you try this: create a new project with the same data folder. See if you still have the same #19 and #32 clusters.
Dear Dr. Chen,
After I saved cluster information, I met some problems with the cluster information. The name of txt file “download_cluster_82.txt” shows that there're 82 papers in the file , but in fact, there are only 21 papers. And some of the papers in this cluster are the same as those in other clusters. However, in the last column of Network Summary Table, the one paper only belongs to one cluster.
Could you be so kind to explain to me why there are only 21 papers and why are the articles in different clusters repeated？ Or if this is not normal, where can I see to solve it?
Thanks
Xin
You need to understand how citing articles and cited references are treated in CiteSpace and what roles they play.
Dear Dr. Chen, thanks very much. I think the 21 papers should be citing articles and they are cited as a role of cited reference. I wonder if this understanding is correct? And shouldn't an article belong to only one cluster? Because you refer that a cocitation network is partitioned to a number of nonoverlapping clusters in your paper "The Structure and Dynamics of Cocitation Clusters:A Multiple-Perspective Cocitation Analysis". I'm so confused. Thanks. Hope to your reply.
See if you can make connections to this figure. When node type is references, each cited reference, which is a member of a cluster, belongs to one cluster only. Its citers, on the other hand, may cite references in multiple clusters. To include citers as references in the network, you would need to use citation expansion when you collect your data.
http://cluster.cis.drexel.edu/~cchen/citespace/papers/citespace2/Figure01.png
fig. 1
fig. 2