본문 바로가기
  • Home

A Study on Developing a Prediction Model of Patent Citation Counts

  • Journal of the Korean Society for Information Management
  • Abbr : JKOSIM
  • 2010, 27(4), pp.239~258
  • DOI : 10.3743/KOSIM.2010.27.4.239
  • Publisher : 한국정보관리학회
  • Research Area : Interdisciplinary Studies > Library and Information Science
  • Received : November 15, 2010
  • Accepted : November 25, 2010
  • Published : December 30, 2010

Yoo, Jae-Bok 1 Young-Mee Chung 2

1한국원자력연구원
2연세대학교

Accredited

ABSTRACT

The purpose of this study is to develop a prediction model of patent citation counts based on major factors which affect patent citation. To this end, we performed multiple regression analysis between the patent citation counts and five explanatory variables such as the number of pages, the number of claims, the reference-average-citation rate, the strength of bibliographic coupling, and the document similarity proved as having 5% or more standardized variances(r2) with patent citation counts, with a test dataset of U.S. patents in five subject fields. As a result, our prediction models showed 58.3% to 89.6% predictability depending on subject fields and revealed the document similarity has the highest impact on citation counts among the five predictive variables in all the subject fields. The result of comparison between the predicted citation counts and the actual ones confirmed the usefulness of the citation prediction models built for each subject field.

Citation status

* References for papers published after 2023 are currently being built.