Siluo Yang

An empirical study on the utilization of web academic resources in humanities and social sciences based on web citations

In this era of a rapid change in the way people finding and using information resources, despite that the academic communication and using patterns for people in the traditional print environment have been studied for many years, the Internet media presents a new and relatively unexplored area for such study. In this article, we explored the distribution and utilization of web recourses in humanities and social sciences based on web citations. We collected 1,421,731 citations listed in 148,172 articles from 493 journals published during the period of 2006–2007 in the CSSCI, which resulted in 44,973 web citations. We counted the amount and types of web resources used in various disciplines, analyzed the URLs frequency from the host-level, fitted the frequency distribution into the regression models with SPSS, and perform the disciplines coupling analysis based on the web citations. We found out that: (a) The distributions of web citations by years or by websites and webpage types are selective and regular; (b) Great disparity exists among various disciplines in terms of using web information, and the high-frequency websites; (c) The frequency distribution of web citations is similar to the Garfield’s citation distribution curve; (d) Some relationships between disciplines are detected, based on the utilization of web information.

Visualizing the intellectual structure of information science (2006–2015): Introducing author keyword coupling analysis

We introduce the author keyword coupling analysis (AKCA) method to visualize the field of information science (2006–2015). We then compare the AKCA method with the author bibliographic coupling analysis (ABCA) method in terms of first- and all-author citation counts. We obtain the following findings: (1) The AKCA method is a new and feasible method for visualizing a disciplines structure, and the ABCA and AKCA methods have their respective strengths and emphases. The relation within the ABCA method is based on the same references (knowledge base), whereas that within the AKCA method is based on the same keywords (lexical linguistic). The AKCA method appears to provide a less detailed picture, and more uneven sub-areas of a discipline structure. The relationships between authors are narrow and direct and feature multiple levels in AKCA. (2) All-author coupling provides a comprehensive picture; thus, a complete view of a discipline structure may require both first- and all-author coupling analyses. (3) Information science evolved continuously during the second decade of the World Wide Web. The KDA (knowledge domain analysis) camp became remarkably prominent, while the IR camp (information retrieval) experienced a further decline in hard IR research, and became significantly smaller; Patent analysis and Open Access emerged during this period. Mapping of Science and Bibliometric evaluation also experienced substantial growth.

Visualizing information science: Author direct citation analysis in China and around the world

Author direct citation analysis (ADCA, also called inter-citation or cross citation) is a new feasible and applicable technique for exploring knowledge communication and discovering scientific structure. This study explored ADCA among prolific, highly cited, and core authors in information science in China and around the world. The results revealed the following. (1) The datasets in China and around the world cover overlapping, but also unique topics. Research subjects on information science around the world can be divided into three categories and 10 clusters; meanwhile, that in China can be divided into three categories and 9 clusters. Chinese scholars who are mostly involved in cross subjects and multi-fields are not as specialized and profound as foreign scholars. An obvious imbalance exists in the evolution of discipline structure around the world, indicating the necessity of a synchronous promotion of research specialty and cross comprehensiveness. Chinese scholars concentrate more on topics such as competitive intelligence, information resource management, and information retrieval, and they focus less on information security and user analysis. (2) Knowledge communication between active authors is stronger than the knowledge flow from highly influential authors to active authors around the world; meanwhile, Chinese researchers tend to adopt the knowledge of authoritative literature. The knowledge flow through bidirectional direct citation is related to mutual knowledge communication. Authoritative scholars are produced when prolific authors cite highly cited authors. The level of mutual recognition among Chinese scholars has not reached that among foreign scholars; in the former, less bidirectional flow of knowledge is involved, and unidirectional flow is limited to geographical proximity, cooperation, or teacher–student relationship. (3) In contrast to traditional author co-citation analysis (ACA), ADCA pays more attention to the mutual interaction among currently active scholars and to mainly showing the current research focus.

Breadth and depth of citation distribution

Abstract This study proposes a new 4D (i.e., spatial, temporal, breadth, and depth) framework for citation distribution analysis. The importance and differences in the breadth and depth of citation distribution are analyzed. Easily computable indices, X , Y , and XY , are proposed, which provide estimates of the breadth and depth of citation distribution. A knowledge unit can be an article, author, institution, journal, or a set of something. Index X , which represents the breadth of citation distribution, is the number of different knowledge units that cite special knowledge units. Index Y , which represents the depth of citation distribution, is the maximum number of citations among several knowledge units that refer to specific knowledge units. Index XY , which synthetically represents Indices X and Y , the feature and focus impacts of a knowledge unit, is index X divided by index Y . We analyze empirically the citation and reference distributions of 84 journals from the “Information science and library science” category of the Journal Citation Reports (2012) at the journal-to-journal level. Indices X , Y , and XY reflect the actual breadth and depth of citation distribution. Differences exist among Indices X , Y , and XY . Differences also exist between these indices and other bibliometric indicators. These indices cannot be replaced by existing bibliometric indicators. Specifically, the absolute values of indices X and Y are good supplements to existing bibliometric indicators. However, index XY and the relative values of Indices X and Y represent new aspects of bibliometric indicators.

The relationship between the author byline and contribution lists: a comparison of three general medical journals

The author byline is an indispensable component of a scientific paper. Some journals have added contribution lists for each paper to provide detailed information of each author’s role. Many papers have explored, respectively, the byline and contribution lists. However, the relationship between the two remains unclear. We select three prominent general medical journals: Journal of the American Medical Association (JAMA), Annals of Internal Medicine (Annals), and PLOS Medicine (PLOS). We analyze the relationship between the author byline and contribution lists using four indexes. Four main findings emerged. First, the number, forms, and names of contribution lists significantly differed among the three journals, although they adopted the criteria of the International Committee of Medical Journal Editors. Second, a U-shaped relationship exists between the extent of contribution and author order: the participation levels in contribution lists were highest for first authors, followed by last and second authors, and then middle authors with the lowest levels. Third, regarding the consistency between author order in the contribution list and byline, every contribution category has a high consistency in JAMA and Annals, while PLOS shows a low consistency, in general. Fourth, the three journals have a similar distribution for the first authors in the contribution category; the first author in the byline contributes the highest proportion, followed by the middle and second authors, and then the last author with the lowest proportion. We also develop recommendations to modify academic and writing practice: implement structured cross-contribution lists, unify formats and standards of contribution lists, draft the author contribution criteria in the social sciences and humanities, and consider author contribution lists in scientific evaluation.

Global science discussed in local altmetrics: Weibo and its comparison with Twitter

Local altmetrics is currently an integral part of the altmetrics landscape. This paper aims to investigate the characteristics of microblog altmetrics of the Chinese microblog platform, Weibo, to shed light on cultural differences and draw attention to local altmetrics in developing countries. The analysis is based on 4.4 million records provided by Altmetric.com. Data collected are from March 2014 to July 2015. It is found that Weibo users discuss global science, more actively compared with several international altmetrics sources. Statistical results show strong evidence of the immediacy advantage of metrics based on Weibo as well as Twitter and the general altmetrics over citations. Distribution of Weibo altmetrics on the article level, source level and discipline level are highly skewed. Overall, compared with Twitter, Weibo altmetrics present similar distributions, with some minor variations. To better understand how and why Weibo users discuss global scientific articles, the top weiboed articles, sources and disciplines are identified and further explored. Our content analysis shows that the common motivation of scientific weibos is to disseminate or discuss the articles because they are interesting, surprising, academically useful or practically useful. Conclusion of articles is the most frequently mentioned element in scientific weibos. In addition, different from Twitter, Weibo users have a preference for traditional prestigious journals.

A domain keyword analysis approach extending Term Frequency-Keyword Active Index with Google Word2Vec model

In bibliometric research, keyword analysis of publications provides an effective way not only to investigate the knowledge structure of research domains, but also to explore the developing trends within domains. To identify the most representative keywords, many approaches have been proposed. Most of them focus on using statistical regularities, syntax, grammar, or network-based characteristics to select representative keywords for the domain analysis. In this paper, we argue that the domain knowledge is reflected by the semantic meanings behind keywords rather than the keywords themselves. We apply the Google Word2Vec model, a model of a word distribution using deep learning, to represent the semantic meanings of the keywords. Based on this work, we propose a new domain knowledge approach, the Semantic Frequency-Semantic Active Index, similar to Term Frequency-Inverse Document Frequency, to link domain and background information and identify infrequent but important keywords. We adopt a semantic similarity measuring process before statistical computation to compute the frequencies of “semantic units” rather than keyword frequencies. Semantic units are generated by word vector clustering, while the Inverse Document Frequency is extended to include the semantic inverse document frequency; thus only words in the inverse documents with a certain similarity will be counted. Taking geographical natural hazards as the domain and natural hazards as the background discipline, we identify the domain-specific knowledge that distinguishes geographical natural hazards from other types of natural hazards. We compare and discuss the advantages and disadvantages of the proposed method in relation to existing methods, finding that by introducing the semantic meaning of the keywords, our method supports more effective domain knowledge analysis.

Community evolution analysis based on co-author network: a case study of academic communities of the journal of “Annals of the Association of American Geographers”

Academic community evolution reveals the development of scientific collaboration among scientists. These social interactions of researchers can be well reflected by co-author network, making it feasible to investigate academic community through looking into co-author network, and to study community evolution through dynamic co-author network analysis. Existing metrics measure an author’s impact or centrality in co-author network individually, rather than considering the academic community as a whole. Besides, co-authors of a paper usually make different contributions reflected in the name order, which is often ignored in traditional co-author network analysis. Furthermore, attention has been paid mainly on those structure-level characteristics like the small-world coefficient and the clustering coefficient, the content-level characteristics like community, author, and topics, however, are crucial in the understanding of community evolution. To address those problems, we firstly propose a “comprehensive impact index” to evaluate the author in a co-author network by comprehensively considering the statistic-based impact and the network-based centrality. Then the comprehensive index value of all authors in a community is added up to evaluate the community as a whole. Further, a lifecycle strategy is proposed for the community evolution analysis. Taking geography academic community as a pilot study, we select 919 co-authored papers from the flagship journal of “Annals of the Association of American Geographers”. The co-author groups are generated by community detection method. Top three co-author groups are identified through computing with the proposed index and analyzed through the proposed lifecycle strategy from perspective of community structures, member authors, and impacts respectively. The results demonstrate our proposed index and strategy are more efficient for analyzing academic community evolution than traditional methods.

The distribution of Web citations

A substantial amount of research has focused on the persistence or availability of Web citations. The present study analyzes Web citation distributions. Web citations are defined as the mentions of the URLs of Web pages (Web resources) as references in academic papers. The present paper primarily focuses on the analysis of the URLs of Web citations and uses three sets of data, namely, Set 1 from the Humanities and Social Science Index in China (CSSCI, 1998-2009), Set 2 from the publications of two international computer science societies, Communications of the ACM and IEEE Computer (1995-1999), and Set 3 from the medical science database, MEDLINE, of the National Library of Medicine (1994-2006). Web citation distributions are investigated based on Web site types, Web page types, URL frequencies, URL depths, URL lengths, and year of article publication. Results show significant differences in the Web citation distributions among the three data sets. However, when the URLs of Web citations with the same hostnames are aggregated, the distributions in the three data sets are consistent with the power law (the Lotka function).

Mining author relationship in scholarly networks based on tripartite citation analysis

Following scholars in Scientometrics as examples, we develop five author relationship networks, namely, co-authorship, author co-citation (AC), author bibliographic coupling (ABC), author direct citation (ADC), and author keyword coupling (AKC). The time frame of data sets is divided into two periods: before 2011 (i.e., T1) and after 2011 (i.e., T2). Through quadratic assignment procedure analysis, we found that some authors have ABC or AC relationships (i.e., potential communication relationship, PCR) but do not have actual collaborations or direct citations (i.e., actual communication relationship, ACR) among them. In addition, we noticed that PCR and AKC are highly correlated and that the old PCR and the new ACR are correlated and consistent. Such facts indicate that PCR tends to produce academic exchanges based on similar themes, and ABC bears more advantages in predicting potential relations. Based on tripartite citation analysis, including AC, ABC, and ADC, we also present an author-relation mining process. Such process can be used to detect deep and potential author relationships. We analyze the prediction capacity by comparing between the T1 and T2 periods, which demonstrate that relation mining can be complementary in identifying authors based on similar themes and discovering more potential collaborations and academic communities.