6

I have a couple of ideas in mind for using machine learning to look for correlations between text and metadata of scientific papers and the number of times the papers have been cited, for which I really need a list of papers and the number of citations.

A Google search easily finds e.g. the top five papers by number of citations, but has not turned up substantially more than that. Is there available, a list of the top, say, 1000 papers by number of citations, or the top N for some reasonably large N?

2

1 Answer 1

8

If you are affiliated at a university, you most likely have access to a subscription-based database of scientific publications like Web of Science or Scopus.

I will show a (simplified) example of how to find the desired result with Web of Science, and below with the free website LENS. But you could do it analogously in Scopus or Dimensions or other related databases.

Web of Science

  • First, narrow down your sample, for example, by publication years 2000 to 2020:

Database search in WoS, 2000-2020

  • Click on "Search".

  • Choose "Sort by Citations: highest first"

You then see the list of all publications from 2000 to 2020, sorted by citation counts. Here are the top 3:

enter image description here

You can then download the results in a *.CSV-file or so, and filter out the first 1000 results.


LENS

If you do not have an affiliation at an university, and therefore no access to commercial databases, there are free alternatives.

LENS, for example, offers such a scholarly search. Here is the analogous query for LENS.

Here is a screenshot of the top 4 papers in LENS:

enter image description here

You will notice that the results differ from that of Web of Science. This is so because every database indexes different kinds of sources, and they count citations differently.

You must log in to answer this question.

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