Research On Visualization Analysis Of Discipline Knowledge

Under the background of big data, networking, visualization, disciplinary integration, ever-changing techniques and the current status of science mapping, the paper integrated theories and methodologies of science of science, network science, visualization and data science to systematically and innovatively study the structure and its evolution of discipline knowledge. In terms of the principle of "a combination of theory and practice", the paper conducted the research from the aspects of theoretical basis, theoretical contents and empirical study. Apart from the introduction and conclusion, the paper consisted of five chapters, which were detailed discussed as follows.Chapter One constructed the theoretical basis of science mapping in terms of the theories of science of science, network science, visualization and data science. Firstly, theories of science of science were detailed discussed from the aspects of scientific revolution structure theory, invisible college theory, conceptual revolution theory and scientometrics theory (the perspective of philosophy, history, society and quantity). Then theories of network science were detailed discussed from the aspects of complex network theory and social network theory (the perspective from general to specific). Then theories of visualization were detailed discussed from the aspects of visual perception and cognitive theory, information visualization theory, visual analytics theory and visualization process theory. Finally, theories of data science were detailed discussed from the aspects of definition, relationships with other disciplines and its contents.Chapter Two constructed the theoretical contents of science mapping from the aspects of related concepts, types, characteristics and methodology. Firstly, it combed the relevant concepts of visualization (scientific visualization, data visualization, information visualization, knowledge visualization and visual analytics) and proposed the concepts of visualization analysis method, knowledge visualization analysis and visualization analysisof discipline knowledge (or science mapping). Then it classified the types of science mapping in terms of the targets of science mapping, visual channels, dimensions, levels and granularities of discipline knowledge units. Then it summarized the characteristics of science mapping, which were "intuitive and efficient","informative","aesthetic" and "novel". Finally, it elaborated the methodological issues which were the analysis mode, methods, tools, and common software combinations of science mapping.Chapter Three classified the discipline knowledge structure into five types (disciplinary cooperation structure, disciplinary theme structure, disciplinary citation structure, interdisciplinary and disciplinary intellectual structure) in terms of the idea of "knowledge production, knowledge content, knowledge transfer and integration" and elaborated the contents of every type from the core parts of science mapping mode, which were the selection, relationship building and visualization of different knowledge units. Firstly, disciplinary cooperation structure analysis was conducted by selection of literature authors (its affiliation, city and country), relationship building of co-authorship and visual analysis of co-authorship network. Then disciplinary theme structure analysis was conducted by selection of literature terms (from titles, abstracts, identifiers and descriptors), relationship building of word co-occurrence and visual analysis of word co-occurrence network. Then analysis of disciplinary citation structure was conducted by selection of citation locations in the full text, relationship building of citing, citation and citing context and visual analysis of citationtotal numbers, citing locations, proximity of citation and citing context. Then interdisciplinary analysis was conducted by selection of citing articles, authors, references and there journals, relationship building of citation co-occurrence and visual analysis of citation co-occurrence network. Finally, disciplinary intellectual structure analysis was conducted by selection of literature authors and cited authors, relationship building of co-authorship and co-citation and visual analysis of co-authorship and co-citation network.Chapter Four classified the evolution of discipline knowledge structure into five types (evolution of disciplinary cooperation structure, evolution of disciplinary theme structure, knowledge base, research front and disciplinary prediction) in terms of the idea of "knowledge production, knowledge content, knowledge transfer and integration" and elaborated the contents of every type from the core parts of science mapping mode, which were the selection, relationship building and visualization of different knowledge units. The differences between the structures and its evolution of science mapping were the latter one contained the factor of time. Firstly, evolution of disciplinary cooperation structure analysis was conducted by selection of literature authors (its affiliation, city and country), relationship building of co-authorship and evolutionary visual analysis of co-authorship network. Then evolution of disciplinary theme structure analysis was conducted by selection of literature terms (from titles, abstracts, identifiers and descriptors), relationship building of word co-occurrence and evolutionary visual analysis of word co-occurrence network. Then knowledge base analysis was conducted by selection of references and cited authors, relationship building of co-citation and visual analysis of co-citation network. Then research front analysis was conducted by selection of citation and term, relationship building of citation term frequency and its co-occurrence and visual analysis of citation network, term frequency evolution and co-word network. Finally, disciplinary prediction analysis was conducted by selection of bibliographic data and full text, relationship building of citing, citation, co-occurrence and so on and visual analysis of latent cooperation prediction, latent prominent authors, latent highly-cited papers and latent disciplinary topics.Chapter Five focused on the empirical research of citation analysis domain from the aspects of disciplinary cooperation and its evolution, disciplinary theme and its evolution, interdisciplinary, disciplinary intellectual structure, knowledge base and research front. It was the practice of the theories above and gave a clue for the latter study of science mapping.The dissertation is one of the achievements and components of the major program "Semantic-based Deep Integration and Visualization of Library Resources"(Code NO.11&ZD152) granted by the National Social Science Foundation of China and the program "Theory and Key Technology of Semantic-based Deep Integration of Library Resources"(Code NO.13YJA870023) granted by the MOE Humanity&Social Science Foundation of China.