What is Special about Patent Citations? Differences between Scientific and Patent Citations

The emergence of pattent bibliometrics as a new branch of scientometrics necessitates a deeper understanding of the relationship between patents and papers. As this connection is established through the linkage between patents and research papers, one must have a clear idea of similarities and differences between patent and paper citations. This paper will investigate to what extent one can not only apply bibliometric methods to patents but also extend the existing interpretative framework for citations in research papers to the field of patent citations. After pointing out some parallels in the debates about the nature of citations in patents and scientific articles, the paper outlines those parts of bibliometric theory covering scientific citations that could be relevant to patent citations too. Then it highlights the specialties and peculiarities of patent citations. One major conclusion is that the general nature of a common framework for both scientific and patent citations would severely limit its usefulness, but research on academic citations might still be a great source of inspiration to the study of patent citations.     

Technological Relevance of Science: An Assessment of Citation Linkages between Patents and Research Papers

Patent citations to the research literature offer a way for identifying and comparing contributions of scientific and technical knowledge to technological development. This case study applies this approach through a series of analyses of citations to Dutch research papers listed on Dutch-invented and foreign patents granted in the US during the years 1987–1996.First, we examined the general validity and utility of these data as input for quantitative analyses of science-technology interactions. The findings provide new empirical evidence in support of the general view that these citations reflect genuine links between science and technology. The results of the various analyses reveal several important features of industrially relevant Dutch science: (1) the international scientific impact of research papers that are also highly cited by patents, (2) the marked rise in citations to Dutch papers on foreign-invented patents; (3) the large share of author-inventor self-citations in Dutch-invented patents; (4) the growing relevance of the life sciences, (5) an increase in the importance of scientific co-operation. We also find significant differences between industrial sectors as well as major contributions of large science-based multinational enterprises, such as Philips, in domestic science-technology linkages.The paper concludes by discussing general benefits and limitations of this bibliometric approach for macro-level analysis of science bases in advanced industrialised countries like the Netherlands.     

Scientific research competitiveness of world universities in computer science

This article evaluates the scientific research competitiveness of world universities in computer science. The data source is the Essential Science Indicator (ESI) database with a time span of more than 10 years, from 01/01/1996 to 08/31/2006. We establish a hierarchical indicator system including four primary indicators which consist of scientific research production, influence, innovation and development and six secondary indicators which consist of the number of papers, total citations, highly cited papers, hot papers, average citations per paper and the ration of highly cited papers to papers. Then we assign them with proper weights. Based on these, we obtain the rankings of university and country/territory competitiveness in computer science. We hope this paper can contribute to the further study in the evaluation of a certain subject or a whole university.     

A multi-view method of scientific paper classification via heterogeneous graph embeddings

The classification task of scientific papers can be implemented based on contents or citations. In order to improve the performance on this task, we express papers as nodes and integrate scientific papers’ contents and citations into a heterogeneous graph. It has two types of edges. One type represents the semantic similarity between papers, derived from papers’ titles and abstracts. The other type represents the citation relationship between papers and the journals or proceedings of conferences of their references. We utilize a contrastive learning method to embed the nodes in the heterogeneous graph into a vector space. Then, we feed the paper node vectors into classifiers, such as the decision tree, multilayer perceptron, and so on. We conduct experiments on three datasets of scientific papers: the Microsoft Academic Graph with 63,211 scientific papers in 20 classes, the Proceedings of the National Academy of Sciences with 38,243 scientific papers in 18 classes, and the American Physical Society with 443,845 scientific papers in 5 classes. The experimental results on the multi-class task show that our multi-view method scores the classification accuracy up to 98%, outperforming state-of-the-arts.

     

Analysis of propensity to patent and science-dependence of large Japanese manufacturers of electrical machinery

Summary The paper aims to clarify the extent to which the results of scientific-oriented research conducted by corporations are reflected in their application-oriented research. Focusing on large Japanese manufacturers of electrical machinery, the paper analyses firm-level data on presentations of scientific papers that represent the results of scientific-oriented research activities, citations of scientific papers in patents, and inventions. The electrical machinery industry, a prototypical science-based industry, has been placing a growing emphasis on scientific-oriented research during the 1990's as is evident from trends in R&D expenses, scientific papers, and inventions. Regression analysis results suggest a complementary relationship between citations of basic scientific knowledge as presented in scientific papers on the one hand and acts of invention on the other hand, in the sense that a rise in citations corresponds to a rise in inventions. Moreover, the results suggest that invention efficiency (number of patent claims per unit of R&D expenditure) has been increasing during the 1990's. Furthermore, the results suggest that, given the exogenous influences on the patent system in Japan, it is necessary to include the number of patent claims when attempting to measure corporate technology development activity through the volume of patent applications. However, there was no finding of a clear relationship between the number of scientific papers and inventions. Implications of these results for corporate R&D strategy are examined.     

A macro study of self-citation

This study investigates the role of self-citation in the scientific production of Norway (1981-1996). More than 45,000 publications have been analysed. Using a three-year citation window we find that 36% of all citations represent author self-citations. However, this percentage is decreasing when citations are traced for longer periods. We find the highest share of self-citation among the least cited papers. There is a strong positive correlation between the number of self-citations and the number of authors of the publications. Still, only a minor part of the overall increase in citation rates that can be found for multi-authored papers is due to self-citations. Also, the share of self-citation shows significant variations among different scientific disciplines. The results are relevant for the discussion concerning use of citation indicators in research assessments. This revised version was published online in August 2006 with corrections to the Cover Date.     

The diffusion of scientific publications: The case of Econometrica, 1987

This paper documents that salient features of (time series of annual) citations to scientific publications might be captured by a Bass type diffusion model. This is particularly useful as it allows for a comparison of these features across journals, across disciplines and over time. For the illustrative case of Econometrica 1987, it is found that the peak in citations occurs at 6.5 years, on average. Also, it is found that after 14 years there is only a little gap between cumulative citations and the estimated total cumulative amount, suggesting that on average the impact of these articles lasts for about 15 years or so. Finally, it appears that these features can partly be explained by the size of the articles, as it is found that longer papers get more citations and peak later.     

Disciplinary determinants of bibliometric impact in Danish industrial research: Collaboration and visibility

Firms are increasingly dependent on networks and network visibility for innovation. Bibliometric impact can be regarded as a measure of a firm's visibility in knowledge-producing networks and may explain why companies publish their results. However, this visibility varies across disciplines. This paper examines publications produced by Danish companies in 1996, 1998 and 2000 to show how citation and collaboration patterns relate in different disciplines. The main findings are that for disciplines characterized by international collaboration and many authors per paper, international collaboration results in a greater number of citations. National collaboration does not, however, seem to make any difference to citation impact in industrial research. In disciplines where multinational collaboration and multi-authorship is uncommon, no clear picture of impact patterns can be obtained. By extension, this research may provide knowledge on how citations of papers in scientific journals can be used as a potential window to scientific networks for firms. This revised version was published online in June 2006 with corrections to the Cover Date.     

Positive and negative aspects of citation indices and journal impact factors

The Hirsch citation index h is nowadays the most frequently used numerical indicator for the performance of scientists as reflected in their output and in the reaction of the scientific community reflected in citations of individual contributions. A few of the possible improvements of h are briefly reviewed. Garfield??s journal impact factor (IF) characterizes the reaction of the scientific community to publications in journals, reflected in citations of all papers published in any given journal during the preceding 2?years, and normalized against all citable articles during the same period. Again, a few of the possible improvements or supplements of IF are briefly reviewed, including the journal-h index proposed by Braun, Gl?nzel, and Schubert. Ascribing higher weighting factors to citations of individual papers proportionally to IF is considered to be a misuse of useful numerical indices based on citations. At most, one could turn this argument on its head and one can find reasons to ascribe an inverse proportionality relative to IF for individual citations: if a paper is considered worthy to be cited even if it was published in a low-IF journal, that citation ought to be worth more than if the citation would have been from a higher-impact journal. A weight factor reflecting the prestige of the citing author(s) may also be considered.     

Exploiting Structural Similarities to Classify Citations

Citations play an important role in the scientific community by assisting in measuring multifarious policies like the impact of journals, researchers, institutions, and countries. Authors cite papers for different reasons, such as extending previous work, comparing their study with the state-of-the-art, providing background of the field, etc. In recent years, researchers have tried to conceptualize all citations into two broad categories, important and incidental. Such a categorization is very important to enhance scientific output in multiple ways, for instance, (1) Helping a researcher in identifying meaningful citations from a list of 100 to 1000 citations (2) Enhancing the impact factor calculation mechanism by more strongly weighting important citations, and (3) Improving researcher, institutional, and university rankings by only considering important citations. All of these uses depend upon correctly identifying the important citations from the list of all citations in a paper. To date, researchers have utilized many features to classify citations into these broad categories: cue phrases, in-text citation counts, and metadata features, etc. However, contemporary approaches are based on identification of in-text citation counts, mapping sections onto the Introduction, Methods, Results, and Discussion (IMRAD) structure, identifying cue phrases, etc. Identifying such features accurately is a challenging task and is normally conducted manually, with the accuracy of citation classification demonstrated in terms of these manually extracted features. This research proposes to examine the content of the cited and citing pair to identify important citing papers for each cited paper. This content similarity approach was adopted from research paper recommendation approaches. Furthermore, a novel section-based content similarity approach is also proposed. The results show that solely using the abstract of the cited and citing papers can achieve similar accuracy as the state-of-the-art approaches. This makes the proposed approach a viable technique that does not depend on manual identification of complex features.     

Slow reception and under-citedness in climate change research: A case study of Charles David Keeling,discoverer of the risk of global warming

The Keeling curve has become a chemical landmark, whereas the papers by Charles David Keeling about the underlying carbon dioxide measurements are not cited as often as can be expected against the backdrop of his final approval. In this bibliometric study, we analyze Keeling’s papers as a case study for under-citedness of climate change publications. Three possible reasons for the under-citedness of Keeling’s papers are discussed: (1) The discourse on global cooling at the starting time of Keeling’s measurement program, (2) the underestimation of what is often seen as “routine science”, and (3) the amount of implicit/informal citations at the expense of explicit/formal (reference-based) citations. Those reasons may have contributed more or less to the slow reception and the under-citedness of Keeling’s seminal works.     

Are multi-authorship and visibility related? Study of ten research areas at Carlos III University of Madrid

Opinions in the literature on the possible relationship between co-authorship and number of citations vary. This paper contributes to the debate with a further analysis of the subject, taking account of the number and quality of citations found for multi-(author, institution, country) and single-authored papers. The study is based on the scientific production of ten Carlos III University of Madrid departmental areas between 1997 and 2003 as reflected in the ISI Web of Science, and the number of times the respective papers were cited between 1997 and 2004. Univariate multifactorial analysis of variance (ANOVA) was used to verify the relationship between multi-authorship and visibility. The correlation between multi-institutional and multi-national authorship and the quartile of the citing journals was analyzed with correspondence analysis. The results show that while multi-institutional and multi-national authorship raise the number of citations, co-authorship and number of citations are unrelated. Correspondence analysis failed to show any correlation between the quartile of the citing journal and multi-institutional or multinational authorship, but did reveal a relationship between citing journal quartile and departmental area.     

Does cumulative advantage affect collective learning in science? An agent-based simulation

Agent-based simulation can model simple micro-level mechanisms capable of generating macro-level patterns, such as frequency distributions and network structures found in bibliometric data. Agent-based simulations of organisational learning have provided analogies for collective problem solving by boundedly rational agents employing heuristics. This paper brings these two areas together in one model of knowledge seeking through scientific publication. It describes a computer simulation in which academic papers are generated with authors, references, contents, and an extrinsic value, and must pass through peer review to become published. We demonstrate that the model can fit bibliometric data for a token journal, Research Policy. Different practices for generating authors and references produce different distributions of papers per author and citations per paper, including the scale-free distributions typical of cumulative advantage processes. We also demonstrate the model’s ability to simulate collective learning or problem solving, for which we use Kauffman’s NK fitness landscape. The model provides evidence that those practices leading to cumulative advantage in citations, that is, papers with many citations becoming even more cited, do not improve scientists’ ability to find good solutions to scientific problems, compared to those practices that ignore past citations. By contrast, what does make a difference is referring only to publications that have successfully passed peer review. Citation practice is one of many issues that a simulation model of science can address when the data-rich literature on scientometrics is connected to the analogy-rich literature on organisations and heuristic search.     

Does it pay to cooperate? A bibliometric case study in molecular biology

Various bibliometric studies report that multiinstitutional or multinational authored papers are more frequently cited than papers that come from a single institute. The conclusion, however, that there is a systematic improvement of scientific success by cooperation on every level of scientific research in leading or mediocre research institutes might be misleading: In a citation analysis of 13 well-known research institutes in molecular biology there was no difference in the average citations per paper with regard to cooperations. In a subsample of 7 German institutes that difference found could be explained by selfcitations. In another case, all articles of a two year sample of an excellent journal in molecular biology, the EMBO-Journal, the same phenomenon could be observed: Differences in the average citations per article with regard to cooperations could be explained by selfcitations.     

Medical research in South Africa: a scientometric analysis of trends,patterns, productivity and partnership

Being a scientifically active country in Africa, South Africa has made significant strides in the production of scientific publications. Medicine is one branch of science that has achieved a remarkable position in this regard. Extracting and analyzing medical publications for three decades and at regular intervals (1975–2005) from the SCI database, this paper pioneers an attempt to find out whether the reported pace of growth in the production of scientific papers in medicine is an effect of partnerships that scholars have with their counterparts within the organization, within the country, or with those in other countries. This paper also presents the unique patterns of scientific research in medicine, taking into account factors such as the count and fractional count of papers, citations, trends of growth, sectoral participation, partners, and publication outlets, and seeks to provide new insights into the directions medical science is taking in South Africa today.     

Patents cited in the scientific literature: An exploratory study of 'reverse' citation relations

This paper reports on a new approach to study the linkage between science and technology. Unlike most contributions to this area we do not trace citations of scientific literature in patents but explore citations of patents in scientific literature. Our analysis is based on papers recorded in the 1996-2000 annual volumes of the CD-Edition of Science Citation Index (SCI) of the Institute for Scientific Information (ISI) and patent data provided by the US Patent and Trademark Office. Almost 30,000 US patents were cited by scientific research papers. We analysed the citation links by scientific fields and technological sectors. Chemistry-related subfields tended to cite patents more than other scientific area. Among technological sectors, chemical clearly dominates followed by drugs and medical patents as the most frequently cited categories. Further analyses included a country-ranking based on inventor-addresses of the cited patents, a more detailed inspection of the ten most cited patents, and an analysis of class-field transfers. The paper concludes with the suggestions for future research. One of them is to compare our 'reverse' citation data with 'regular' patent citation data within the same classification system to see whether citations occur, irrespectively of their directionality, in the same fields of science and technology. Another question is as to how one should interpret reverse citation linkages.     

An impact-citations-exergy (<Emphasis Type="Italic">iCX</Emphasis>) trajectory analysis of leading research institutions in India

A thermodynamic analogy allows bibliometric research assessment of information production processes to be based on a scalar indicator which is an energy-like term called exergy. Derived from standard indicators like impact, citations and number of papers, the exergy indicator X is a multiplicative product of quality and quantity of a scientist’s or group’s performance using available bibliometric information. Thus, given the bibliometric sequences of leading research agencies and institutions, research performance can be displayed as trajectories on a two-dimensional map as time progresses. In this paper, we track the performance of several of the leading players contributing to academic scientific research in India.     

The effect of keyword repetition in abstract and keyword frequency per journal in predicting citation counts

This paper investigates an association between two new variables and citations in papers. These variables include the abstract ratio (the sum of repetition of keywords in abstract divided by abstract length) and the weight ratio (the frequency of paper’s keyword per journal). The data consist of 5875 papers from 12 journals in education: three journals from each SCImago quartile. The researchers used semi-continuous regression to model the data and measure the impact of the proposed variables on citations. The results revealed that both abstract ratio and weight ratio are statistically significant predictors of citations in scientific articles in education.     

Motivations for URL citations to open access library and information science articles

Summary We define the URL citations of a Web page to be the mentions of its URL in the text of other Web pages, whether hyperlinked or not. The proportions of formal and informal scholarly motivations for creating URL citations to Library and Information Science open access journal articles were identified. Five characteristics for each source of URL citations equivalent to formal citations were manually extracted and the relationship between Web and conventional citation counts at the e-journal level was examined. Results of Google searches showed that 282 research articles published in the year 2000 in 15 peer-reviewed LIS open access journals were invoked by 3,045 URL citations. Of these URL citations, 43% were created for formal scholarly reasons equivalent to traditional citations and 18% for informal scholarly reasons. Of the sources of URL citations, 82% were in English, 88% were full text papers and 58% were non-HTML documents. Of the URL citations, 60% were text URLs only and 40% were hyperlinked. About 50% of URL citations were created within one year after the publication of the cited e-article. A slight correlation was found between average numbers of URL citations and average numbers of ISI citations for the journals in 2000. Separating out the citing HTML and non-HTML documents showed that formal scholarly communication trends on the Web were mainly influenced by text URL citations from non-HTML documents.     

The use of deciles of the citation impact to evaluate different fields of research in Israel

One often uses the average citation impact factor in order to perform international comparisons between the levels of scientific performance within given disciplines. In averaging over all (or all cited) papers one may gives undue weight to papers with few citations while, in fact, the standing of a country within a given field would be better assessed by looking only at the “successful” papers in that discipline. The present papers suggests that one should do so by averaging citations only over the ten (or twenty) percent of the most cited papers in a discipline and use these in order to establish a ranking between countries. The case of Israel is used as an illustration of this approach.