I believe that following this change to jcn the formula will be comparable to the DiShIn implementation.
Is there a reference implementation or publication that these similarity formulas should be based off of? One challenge is that I'm seeing different formulas in the literature for these metrics.
I believe that following this change to
jcn
the formula will be comparable to the DiShIn implementation.Is there a reference implementation or publication that these similarity formulas should be based off of? One challenge is that I'm seeing different formulas in the literature for these metrics.