Identifying SCI covered publications within non-patent references in U.S. utility patents

Masashi Shirabe ()
Additional contact information
Masashi Shirabe: Tokyo Institute of Technology

Scientometrics, 2014, vol. 101, issue 2, No 6, 999-1014

Abstract: Abstract In order to evaluate approaches for identifying science citation index (SCI) covered publications within non-patent references (NPRs), the author employs a computer science method that uses two key indicators, recall and precision, to evaluate the relevance of information retrieval systems. There are two primary reasons that this method is adequate: in contrast to the retrievability ratios used previously, first, this method can evaluate two dimensions of matching accuracy, and second, results of its evaluation are independent of the intermediate outcome. The author then proposes an approach for identifying SCI publications within NPRs that consists of five steps: (1) data collection, (2) creation of supervised and test data, (3) selection and execution of matching algorithms, (4) evaluation of algorithms and optimization of their combinations, and (5) evaluation of optimized combinations. A comparison of the proposed and conventional approaches showed that the proposed approach works well, with results far better (99 % precision and 95 % recall) than the target implicitly set in previous studies. The author also applied the approach to comprehensive NPR data in U.S. utility patents registered between 1992 and 2012 and checked the performance. Results showed that the approach could identify SCI publications from within millions of NPRs in an acceptable time (i.e., within a couple of weeks) and that it performs as expected from the evaluation in step 5. On the basis of these results, the proposed approach is considered of value in studies on relations and/or interactions between science publications and patents.

Keywords: Non-patent references; U.S. utility patents; Science citation; Automated matching; 68U15 (search for similar items in EconPapers)
JEL-codes: C81 (search for similar items in EconPapers)
Date: 2014
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
http://link.springer.com/10.1007/s11192-014-1293-8 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:scient:v:101:y:2014:i:2:d:10.1007_s11192-014-1293-8

Ordering information: This journal article can be ordered from
http://www.springer.com/economics/journal/11192

DOI: 10.1007/s11192-014-1293-8

Access Statistics for this article

Scientometrics is currently edited by Wolfgang Glänzel

More articles in Scientometrics from Springer, Akadémiai Kiadó
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().