8
$\begingroup$

I know of the Jaccard index and the Sørensen-Dice coefficient for computing set similarity, but have been unable to find any other algorithms related to set similarity. This site contains quite a few resources for vector similarity, but that's not what I want.

What other set-similarity measures exist?

Improve this question
$\endgroup$
1
  • $\begingroup$ You can treat being in a set as an variable (or being in the first set as a classifier and being in the second as the concept). Using this reduction you can apply all the supervised learning metrics - accuracy, mutual information, etc. $\endgroup$
    – DaL
    Commented Jun 15, 2017 at 7:14

1 Answer 1

Reset to default
6
$\begingroup$

Other measures are:

  • Overlap Coefficient: $\frac{|A \cap B|}{min(|A|,|B|)}$
  • Tversky index: $|A\cap B| + \alpha|A\setminus B| + \beta|B \setminus A|$ where $\alpha$ and $\beta$ are positive numbers.
Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Tversky is the generalization of Jaccard and DIce $\endgroup$
    – A_Arnold
    Commented Apr 18, 2019 at 20:27

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.