I am doing research on clustering evaluation measures in overlapping subspaces clusters. However, I still haven't found something that really helps me to measure the quality of a clustering. So far what I have seen most appropriate to me would be to use the Jaccard index. The figure below shows an example of what I get with overlapping subspace clustering.
Expected result for the clustering.
As you can see, I can see overlap in rows and columns. So, if I use traditional measures (for example ARI, Silhouette, DB) I don't capture the quality taking into account the overlap. Has anyone studied anything in this regard? Sorry if I missed any presentation rules, this is my first time interacting here.