A Study On The Application Of Citation Analysis Visualization In Science History

As one of the most important methods for scientometrics, citation analysis is an effective measure for the study of science structure and history. The documents which are important for the development of one discipline can be identified by studying the longitudinal references connections between papers, and then, both the bibliographic antecedents and descendents of its principal, often primordial papers and authors can also be identified. This is very helpful for the science historians to find where a particular topic began and how the topic developed. Co-citation analysis can identify the author colony of one subject, visualizing knowledge domains, measure the movement of a research paradigm, searching the intellectual turning points, detecting and mapping thematic changes etc. Instead of deriving a salient semantic structure from the huge documents, co-citation analysis emphasizes the unique value of higher order interrelationships between documents or between author through calculate and analysis the citation count. Citation analysis, then, in a certain extent, can make the history research more scientific and more impersonal for the researchers whose special knowledge or talents in history and the subject being researched are not enough.At the present time, visualizing citation analysis which was developed from the combination of information visualization and citation analysis is a research front abroad. But the research of this topic is few in our country. This paper select the topic of hybrid rice as a case, try to explore the methods, steps and technology of application of citation connection visualization in the research of science history in China. The applicability and veracity of citation visualization to science history research is also discussed in this paper. To be more specific, the following aspects of this research are described in this paper.(1) As a preparation of the study, the origin and the development of citation analysis is described, and the interrelation between SCI and citation analysis is also discussed. The three type of citation analysis and its application are summarized respectively, which is citation quantity analysis, citation timeline network analysis, and citation relation analysis. The status of research of citation analysis in our country is presented. Compared to the research overseas, domestic research is still in the beginning stage. The reasons why domestic research mainly focus on the basic aspect of citation analysis are discussed in this paper. The author also introduces the function and features of 5 citation databases which are developed in China in recent years.(2) The status about the research and application of citation visualization abroad is summarized in this paper. The oversea hot research in citation visualization are mainly focused on the following aspects: uncovering an event of importance and its relation to the other key events in the history of a research subject through tracing the bibliographic antecedents and descendents of its principal, often primordial papers and authors; mapping and visualizing the structure of science and the knowledge domain through the visualization of document co-citation, author co-citation, category co-citation, journal inter citation, or even subject co-citation; Describing some software such Pajek, UCINET, VxInsight, KNOT, IN-SPIRE, which can be used to visualizing citation analysis. These kinds of software have been used widely abroad but known a little by domestic researchers; the advantages and disadvantages of a varieties of graph generated from citation visualization are analyzed and discussed. At the end of the chapter, the applicability of citation analysis visualization is discussed and the experiences of applying the citation analysis in the research of science history are presented.(3) On the basis of summarizing the overseas experiences of applying citation visualization in the research of science history, the citation data on the topic of hybrid rice research was downloaded respectively from SCI and Chinese Citation Index Database, and then the data was import into the local ACCESS database. The historiography was constructed through a citation historiography visualization software—HistCite. The key documents which represent the key events and their citation relationship during the development of research on the topic of hybrid rice can be identified. The history flow on the topic of hybrid rice can be interpreted. In order to evaluate the reliability and the accuracy of the citation analysis as a method applied in the historical research, the result of citation analysis was compared with the opinion of some experts whose research topics are hybrid rice. The conclusion is, that citation timeline web analysis is a reliable method in the process of historical research.(4) The methods and the technology of visualizing knowledge domain via co-citation analysis visualization are summarized, analyzed and compared in this paper. 4 co-citation matrixes are constructed via 3 methods to obtain the co-citation raw count . The first is retrieving the selected author co-citation count via online searching from the Chinese Citation Index Database. These co-citation counts are assembled in a matrix which is called "Matrix A". The second is collecting all the documents which cite the documents written by the selected authors and than compute the author co-citation raw count. "Matrix B" was generated from this method. The third is collecting all the source documents on topic of hybrid rice and all the cited documents by these source documents respectively from Chinese Citation Index Database and SCI. The authors co-citation counts are computed according the citation connections between the local documents resources. The matrix which is generated from the data of SCI is called "Matrix D", and the other is "Matrix C". Path Finder Network was used to analyze and process the data matrixes and 4 knowledge domain maps had been created. The authors colony and the subfields which was represented by the authors in the maps were identified and described. The distinguish of the 4 maps and the reasons that causes the distinguish are discussed. A viewpoint was presented in this paper : the more documents are included in a citation index database, the more accurate and reliable it will be for citation analysis. In order to find out the changes of paradigm in the topic of hybrid rice, the timeline visualization of document cluster was also presented in the paper.(5) Although citation analysis visualization had been applied to historical research, , information science, and even management as an important measure method, there are still some factors that affect the reality and veracity when it is applied. The main factors are: the rules of references description in a paper are not obeyed by all the authors; the veracity of citation data can not be ensured because of pool quality of some citation index, mistaken information of reference entries, ignoring co-author while calculating the citation count, and the range of resources in which the citation count is computed etc; because bibliographic citation has been an established convention of scientific publication only since the early part of the twentieth century, beyond all doubt, it limits the application of citation analysis in history-of-science studies to more earlier time.The main contribution of this paper is as follows:(1) The methods and the key technology of visualizing knowledge domain via co-citation analysis visualization are summarized, analyzed and compared in this paper. It is the first time this sort of study had been developed in our country. (2) It is the first time that the types and the methods of citation visualization which is applied in science historical research are analyzed and summarized. It is also the first time that the Chinese citation data was used to visualizing the citation timeline network and generating the Histriograph of hybrid rice research. In this manner the history of hybrid rice studies is showed in the paper. This is a useful experiment for the history of science studies in our country.(3) It is the first time in our country that the Citation Visualization System—HictCite was introduced and evaluated. It can be a important tool for identifying key events, their chronology, relationships, and relative importance in working out the history of a given scientific effort.(4)It is the first time in China that PFNET was applied in knowledge domain visualization and three methods were used to calculate the author co-citation count. The first method is calculating all the co-citation count between author pairs including the co-author. The second is only calculating the co-citation count between the author pairs who are the first authors of papers. The third is calculating the co-citation count only according the citation in a rather small range of documents resource which is about one research subject. Three type of knowledge domain maps were generated from the co-citation count matrixes which is obtained by the above three different methods. The differences of the maps the reasons which cause the differences are analyzed and discussed. The existence of this differences were proved by the experiment which provided a reference for use to co-citation research.