Configuring Scopus (optional)

In order to run ReCiter, users must be set up an import of records from PubMed using the PubMed Retrieval Tool. Scopus, on the other hand, is optional.

Configuring

To set up Scopus, please see these directions in the Scopus Retrieval Tool repository.

Effect on accuracy

Use of Scopus, which depends on a standard license, is optional but helpful. In one experiment, we found that using Scopus data improved algorithm recall by approximately 5% and precision by 0.2%, when compared against using PubMed data alone. Using Scopus data tends to be more useful in cases where affiliation data is not present in the PubMed record or the full name of authors isn't indexed. This is more common in older papers. For that reason, Scopus may offer less of an advantage for more recently published papers.

So long as it's okay with your license, it would be possible to use alternatives to Scopus such as Web of Science. Of course, you would need to figure out how to parse these records, among other things.

Note that Scopus is only used as a compliment to PubMed. In other words, ReCiter only uses identifiers (PMID, DOI) to find additional data (namely affiliation and full name) about a candidate article.

Coverage

Weill Cornell has thus far retrieved 1.595 million articles. Of these, 1.491 million or 93%, match to an equivalent record in Scopus.