초록

Abstract

This study aims at proposing three prestige indices—closeness prestige, input domain, and proximity prestige- as useful measures for the impact of a particular node in citation networks. It compares these prestige indices with other impact indices as it is still unknown what dimensions of impact these indices actually measure. The prestige indices enable us to distinguish the most prominent actors in a directed network, similar to the centrality indices in undirected networks. Correlation analysis and principal component analysis were conducted on the author citation network to identify the differentiated implications of the three prestige indices from the existing impact indices. We selected simple citation counting, h-index, PageRank, and the three kinds of centrality indices which assume undirected networks as the existing impact measures for comparison with the three prestige indices. The results indicate that these prestige indices demonstrate distinct impact dimension from the other impact indices. The prestige indices reflect indirect impact while the others direct impact.

초록

본 연구는 약물 연구 분야에 속하는 특허 사이에 나타나는 지식의 흐름을 살펴보고 이들 간의 영향력을 파악해보기 위해 특허데이터에서 나타나는 인용 관계를 분석하였다. 특허데이터의 수집은 Google Patents에서 진행하였다. 약물 연구와 관련된 특허 문서를 검색하여 상위 25개의 출원인을 선정하였고, 이를 바탕으로 출원인 사이에서의 인용 관계를 알아보고 각 출원인의 각 문서에 대한 피인용빈도와 순위를 활용하여 h-지수와 h-지수의 파생지표들의 값을 계산하여 비교하였다. 분석 결과를 종합하면, ‘Pfizer, MIT, Abbott’ 등의 출원인이 약물 연구 분야에서 영향력이 높은 출원인으로 드러났다. 5개의 계량서지학적 지표 중에서 g-지수와 hS-지수가 서로 유사한 결과를 보여주었고, 총인용빈도, 최대인용빈도, CPP의 순위를 가장 잘 반영하는 지표로 나타났다. 또한, 총인용빈도, CPP, 최대인용빈도 순으로 5개의 계량서지학적 지표와의 상관관계가 높았다. 한편, 기존의 특허 출원인의 기술적 영향력을 나타내는 것으로 알려진 지표인 CPP만으로는 정확한 비교가 어려운 경우도 나타났다.

Abstract

This study analyzes the relationship of citations appearing in the patent data to understand knowledge transfers and impacts between patent documents in the field of pharmaceutical research. Patent data were collected from a website, Google Patents. The top 25 assignees were selected by searching for patent documents related to pharmaceutical research. We identify the citation relationships between assignees, then calculate and compare the values of h-index and derived indicators by using the number of citations and rank for each document of each assignee. As a result, in the case of pharmaceutical research, the assignee, such as ‘Pfizer, MIT, and Abbott’ shows a high impact. Among the five bibliometric indicators, the g-index and hS-index show similar results, and the indicators are the most related to the rankings of Total Citation Frequency, Cites per Patents, and Maximum Citation Frequency. In addition, it is highly related to the five indicators in the order of Total Citation Frequency, Cites per Patents, and Maximum Citation Frequency. In some cases, it is difficult to make an accurate comparison with Cites per Patents alone, which is previously known to indicate the technological influence of patent assignees.