I have to compute a similarity measure between different sets (Actually they are more like maps than sets). A weight is associated to each element of the set.

The sets I want to compare represent different journal or conferences. Each element of the set is an author and the associated weight is the number of papers published in that venue. The goal of my analysis is to discover whether there are similarities in the most prolific authors of different venues.

I was thinking to represent the content of the sets in a vector space (Using the elements as features and the weight as their value) and compute a cosine distance. Is it a good approach or there is something better I should use?

  • $\begingroup$ It's pretty hard to provide good advice without knowing the objective of your analysis or more details about the nature of your data. Would you care to elaborate on these? $\endgroup$ – whuber Apr 24 '12 at 13:13
  • 1
    $\begingroup$ I edited the question to add more details. I hope now it is clearer what I want to do. $\endgroup$ – mariosangiorgio Apr 24 '12 at 13:23
  • $\begingroup$ You could look into correspondence analysis $\endgroup$ – kjetil b halvorsen Jan 17 '19 at 9:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.