■■■ 3.2 Classification based on similarities in the interdisciplinary connections of organizations4.1 Collected Data4.2 Analytical Methods and Results─ 12 ─Fig. 1 Researchers’ and an Organization’s Fields of View■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■the disciplines via researcher A in Fig. 1(b) and researcher B in Fig. 1(c) are clearly shown. In the method for visualizing organizational research power and interdisciplinary integration, the connecting lines are thicker because of the understanding that knowledge sharing between disciplines is enhanced based on the number or ratio of mediating researchers. This chart of the organization’s research capabilities and cross-disciplinary integration can be used for comparison between organizations.Hierarchical cluster analysis method is commonly used for finding subgroups of multivariate data. This method creates a dendrogram based on the similarity of the items analyzed. The researcher can choose where to cut the dendrogram to create clusters. This method does not have a generally accepted stopping rule for researchers to find the best set of clusters (Zupic & Cater, 2015)12). Procedures for hierarchical cluster analysis include single, complete, average linkages, as well as Ward’s method. Of these, Ward’s method is the most frequently used for bibliometric analysis; McCain (1990) stated that both complete linkage and Ward’s method produce similar interpretable results13).Research papers with high similarity in interdisciplinary connections of organizations are gathered and grouped. In the classification process, we conducted a hierarchical cluster analysis using interdisciplinary connections as a variable for each organization and visualized the results using a dendrogram. Ward’s method was used to determine the distance between clusters (Fig. 3). This process is considered effective in simplifying the characteristics of each group when interpreting the results.■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■Note. We used the bibliographic data of Web of science Core collection.The number of studies on big data continuously increased in 2018, to reach 6,054 respectively. We used the bibliographic data of Web of Science (WoS) core collection, which is the one of the biggest bibliographic databases. We have permission to use the Web of Science (WoS) Core Collection, a subset of which Clarivate Analytical Inc. provided to the Institute of Statistical Mathematics. This database has been optimized for bibliometric data analysis; using them, some unavailable items on the regular WoS site are accessible for analysis. However, due to contractual regulations, this database only contains data up to 2018 that we have used in our research.Fig. 2 shows the connections between the research fields of the top 10 countries in terms of the number of big data papers, the top three being China, the United States, and England. China ranked first, with a complete network of chemistry, clinical medicine, and engineering (3‒4‒7); clinical medicine, computer science, and engineering (4‒5‒7); and chemistry, clinical medicine, and molecular biology and genetics (3‒4‒14). The United States ranked second, with clinical medicine, computer science, and engineering (4‒5‒7); chemistry, clinical medicine, and molecular biology and genetics (3‒4‒14); and biology, biochemistry, and chemistry (3‒4‒14). Biology and biochemistry, chemistry, and molecular biology and genetics (2‒3‒14) were complete networks. Clinical medicine and general social sciences (4‒21) were also connected. The United States has a wide range of connections but none was strong. England, in the third place, has numerous connections, mainly in clinical medicine (4) and ■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■ff■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■■4. Analysis
元のページ ../index.html#14