I want to cite the author of nearest centroid classifier. However, I can find neither the author nor the time when this technique was used first.

Of course, I may cite the many book chapters which explain extensively this technique. However, in general, where should I look to find the author's information?

  • 1
    In economics there is the concept of "folk theorem", which is applied to ideas that are well-known but have obscure origin. In that case, usually some influential, useful, or canonical discussion is cited. Perhaps some similar conventions exists in CS?
    – henning
    Jun 7, 2021 at 7:23
  • Does this answer your question? How cite sources of 100-years-old analytical technique in a theory section?
    – GoodDeeds
    Jun 7, 2021 at 10:21
  • 1
    I don't think that's a duplicate, since incud explicitly wants to cite the original author (whether to give credit or provide a sense of history). So this is more a question on how one would trace the roots of an idea down in the literature, than on how one would cite basic knowledge.
    – Anyon
    Jun 7, 2021 at 11:29
  • Thank you, @Anyon is correct
    – incud
    Jun 7, 2021 at 13:32
  • 1
    @henning, the canonical authors for CS are Knuth and Sedgewick for basic algorithms, but for classification I'd start with the authors listed at Wikipedia and drill down from there. Jun 7, 2021 at 21:08

2 Answers 2


I would start by talking with the librarian research specialists at your institution's academic library. They may have a discipline specialist in your field, or they may have a history of science specialist. They are very adept at using the research databases, almost for sure they can get you on tracks you would otherwise not take.


The general procedure is to look at a recent book or survey article that mentions this and see what is cited. If not the original author, then perhaps an earlier article. Then look in there, see what is cited, etc.

The hard part is knowing when to stop. If I cannot find the original author after a few hours of searching I would stop unless I am finding the process interesting and I have the time. It is acceptable to stop after a reasonable length of time and just cite a textbook or survey article.

You need to know your audience. I find many mathematicians love to read about a 150 year old paper in Flemish, while most physicists are not so historically minded. Nobody is right, expect Referee B who asks you to dig further back.

You must log in to answer this question.

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