Theory, Methods And Tools Of Knowledge Visualization And Application Study In Military Medicine

With the rapid development of science and technology, scientific knowledge isexploding. How to effectively extract the visual information about the law ofknowledge development and its structure characteristics from the massiveknowledge, has become an important subject of scientific research. Militarymedicine conforms to the development rules of general science, the rapid growthof military medical knowledge also needs to reveal the development law andstructure features from a large amount of military medical knowledge. Moreover,two types and first-class development targets were proposed in the TwelfthFive-year Plan of the Academy of Military Medical Sciences. According to thefirst-class s development goal, the need for a comprehensive grasp ofthe main research directions of important military medical institutions andtheir change trends, the need for scientific evaluation of disciplines,excellent research teams and excellent talents etc. were proposed. All of theserequirements can be partially satisfied by the knowledge visualization method basedon quantitative analysis.In this study, literature survey, system analysis, comprehensive analysis andexpert consultation were used to combine widely through the development process,basic research, and knowledge visualization research problems and relevance indifferent disciplines of the knowledge visualization theory. The definition, researchcontent and research object of the knowledge visualization in the intelligencecommunity were raised. The classification, the data processing flow of differenttypes, mathematical and computational methods involved in the data processingflow of the knowledge visualization methods were comprehensively reviewed. Theknowledge visualization experimental study on the scientific researchinstitutions, research directions, research disciplines and research teams wereanalyzed with the methods of integrated analysis of knowledgevisualization methods, integrated analysis of multivariate statistical methods, theknowledge visualization tool integrated application and empirical study. And theapplicability and limitations of the selected methods and tools were well proven. The research consists of three parts as follows.Part1：The Theoretical StudyFirstly, the formation and development of the knowledge visualization theoryin the education and intelligence fields were systematically analyzed, knowledgevisualization is defined as the field of knowledge visualization examines the use ofvisual representations to improve the creation and transfer of knowledge between atleast two people in education circle. We found that knowledge visualization ineducation circle lacked visual effect research, the classical characterization methodsused were just some graphical representations, not forming a pattern or practice ofbatch of tools. Knowledge visualization in the intelligence community can besummarized as scientific knowledge displayed by the chart using a varietyof mathematical and computer based methods, to reveal the law of scientificdevelopment and internal structure of scientific knowledge.Secondly, the basic theory of knowledge visualization was summarized asmetrology and mapping knowledge domains. The metrology is classified asstatistical analysis, mathematical models to reveal the laws of scientificdevelopment, and revealing the knowledge structure. Mapping knowledgedomains are a series of softwares, who integrate metrologymethods revealing structure characteristics of knowledge units.Lastly, the definition, research content and research object of the knowledgevisualization in the intelligence community were put forward. Knowledgevisualization in the intelligence community is defined as methods that usingmathematics and computer aided method, with charts and other visual models toreveal the amount, quality, development process and structure relationship ofscience and technology literatures, patents and achievements etc. The researchcontent of knowledge visualization includes measuring knowledge indexes,revealing knowledge development laws and revealing knowledge structurecharacteristics. Ten common science and technology indicators and sevenscientometric indicators were summarized. The research objects of knowledgevisualization include authors, inventors, keywords, subject headings, journals,disciplines, cited authors, cited references, citation journals,organizations, countries, regions and time etc. Part2：The Methods and Tools StudyKnowledge visualization can be divided into three types: the statisticalclass, the mathematical model and the knowledge structure feature. The dataprocessing flow of each one were systematically analyzed, and the flow ofstatistical and mathematical model consists of data acquisition, data processingand data analysis. The flow of knowledge structure feature can be divided intoseven steps: data searching, data preprocessing, selection of knowledge units,knowledge unit correlation, data normalization, data reduction, visualization andinterpretation of results. Mathematical methods during each step were summarized.Among these methods, establishing two-dimensional,three-dimensional or multidimensional co-occurrence of knowledge units is themain method to reveal the social correlation structure, conception structure andknowledge structure etc. of the research objects.The common integrated tools were classified as statistical analysissoftware, social network software and specialized software. The features and use ofeach software were described. TDA, BICOMB, SPSS and CiteSpaceII were chosento analyze the experimental study according to the need for data processing andknowledge visualization of military medicine.Part3：The Application and Empirical StudyFirstly, the knowledge visualization of military medicine was classifiedaccording to purpose, function and meaning. Knowledge visualization of militarymedicine can be classified as knowledge visualization in scientific research andknowledge visualization in management of scientific research according to thepurpose. The function of knowledge visualization of military medicine can beclassified as four types, statistical one, association network one, spatial structureone and route decision making one. The statistical knowledge visualization includesscience and technology indicators, scientometric indicators, the mathematical modelof science development rule, and the statistics of one-dimensional knowledgeunit etc. Association network knowledge visualization is the connection betweenthe attributes of knowledge nodes. The spatial structure one is to reveal the structureof research object in knowledge space from the knowledge position angle. Routedecision making one is a dynamic analysis with time based on the associationnetwork and spatial structure knowledge visualization. The common meaning ofknowledge visualization includes research institutions, research directions, researchfields and research teams. Then, all the military hospitals were chosen as an example of experimentalstudy on knowledge visualization of military medical institutions. Based on the SCIpapers and TI patents, the military hospital research directions and their changeswith time, the cooperation networks among military hospitals and their changeswith time, external cooperation networks and reasons, potential cooperation orcompetition fields among authors or research groups and possibility of rapiddevelopment of potential fields were analyzed. This research proved that wordco-occurrence, institution cooperation, author cooperation, word co-occurrence withtime zone and institution cooperation with time zone can display the knowledgevisualization of military medical institutions. While, compared with the patent map,patent term co-occurrence clustering method is more suitable for the patenttechnology division of important military medical institutions.Then, the research directions of the importantmilitary medical institutions worldwide were chosen as an example of experimentalstudy on knowledge visualization of military medical research directions. The mainresearch directions and their changes with time were analyzed. Comparing with theresults of information investigation, the study proved that clustering andmultidimensional scaling analysis of SCI papers keyword co-occurrence couldmostly reveal the research directions of military medicine. Whereas the data ofnon-English papers and patents should be added to reveal the research directions ofmilitary medicine.Then, the sleep deprivation was chosen as an example of experimental study onknowledge visualization of military medical field. The sub fieldclassification, sub field direction changes with time, field emergence terms,important institution cooperations, important author cooperation and importantresearch foundation were analyzed. This study proved that the MeSH termco-occurrence could classify the sub fields better than keyword co-occurrence, burstterms could reveal unexpected research direction and technology, and referenceco-citation could reveal the important research foundation of the research field.Lastly, the Proteomics Innovation Team from the Academy of Military MedicalSciences was chosen as an example of experimental study on knowledgevisualization of military medical research team. The main research directions, themain members, cooperation with the outside institutions, author cooperation, substructures of the team and the main research foundation of the team were analyzed. This study proved that the author cooperation networks could reveal the innerstructure of the team, the journal co-citation and reference co-citation could revealthe research foundation of the team, H index could evaluate the academic influenceof research team and researchers. But the quality evaluation indexes, such as theevaluation indexes of journals, peer-review scores (such as f1000score) ofliteratures and patent intensity index (such as the Innography patent intensity score)should be added.