0

I've been pondering a lot about this whole citation business. Does anyone here have ideas on how one would measure the quality of different citations? Objective answers only, of course.

4
  • 3
    What do you mean by quality of citation? How good is the citation Or the qualitative features of the citation (e.g., support, critique, acknowledgement, etc.)? And if you mean "how good", what do you mean by good (coming from reputable or high impact source or something else)? Jun 23, 2021 at 6:25
  • 1
    1. The qualitative features of the citation. 2. Coming from reputable (not high impact because I think that's a silly measure). The goal would be to give rewards based on citations over time to the creators of research, and I'm wondering how to sort for quality. Jun 23, 2021 at 6:29
  • Do you mean automatically? It would be possible for an expert to make such an evaluation with a lot of work and judgement.
    – henning
    Jun 23, 2021 at 6:55
  • Yes, automatically, if possible. Eg. Should there be weighting for quality if referenced in a paper that has high citations, or is rated high quality by experts at publication. Jun 23, 2021 at 7:11

1 Answer 1

1

In essence, you want to weigh citations higher that come from papers that are themselves "important" in some way. If you think of papers as nodes of graphs and citations as (directed) edges, then you are asking questions such as the "centrality" of a node in that graph.

In some sense, this is also related to the PageRank algorithm google uses/used to use: They also want to rank pages higher that don't just have lots of links to them, but links from other important pages.

2
  • What method do you think is most relevant today for measuring importance of citations in a graph network? Von Neumann, HITS? Jun 25, 2021 at 1:39
  • I don't know. There are many centrality definitions in graph theory and the community does not seem to have come to a conclusion which is "the right" one to consider. Jun 26, 2021 at 4:31

Not the answer you're looking for? Browse other questions tagged .