應用網路模式比較演算法分析台灣醫院作者學術合作的情形（自2022到2023年）

在書目計量學中，網絡分析對於解讀龐大的學術資料集至為重要，但傳統方法尚未比較網絡間的相似性與差異。本研究介紹一具為書目計量資料量身定制的網絡圖騰比較演算法，用以對比2022年至2023年台灣醫院學者撰寫的期刊論文中，基於研究機構、部門與作者的主體網絡。我們評估了2022年至2023年收錄在科學引文索引（SCI）的11,217篇與9,189篇論文。開發這具網絡圖騰比較算則（NPCA），利用作者合作地圖的相似性係數，來比較作者合作主體在綢絡上的相似「生與差異。係數分為相同（大於0.7）、相似（0.5至0.7之間）、不同（0.3至0.5之間）、及截然不同（低於0.3）。結果顯示，這兩年作者合作主體的網絡圖騰相似性係數分別為0.66、0.83、和0.28，說明部門存在相同的模式（大於0.70）、研究機構相似（大於0.5）。而於第一作者和通訊作者的作者間，在2022年和2023年之間，發現了截然不同的綱絡圖騰（小於0.3）。NPCA有效地分析書目計量的網絡圖騰，突顯了這兩年部門和醫院機構作者間合作主體的一致趨勢，但作者本身的網絡圖騰則有顯著的變化。這為書目計量比較網絡圖騰，鋪出新方向與渠路，也可於未來應用於研究機構、部門與作者的主體網絡之外，例如醫院團隊資源管理的網絡圖騰比較。

In the field of bibliometrics, network analysis is pivotal for interpreting complex academic data sets. However, traditional methods have not fully explored the comparison of network similarities and differences. This study introduces a novel network-pattern comparison algorithm (NPCA) specifically tailored for bibliometric data. It applies this algorithm to analyze networks based on institutes, departments, and authors (IDA) derived from journal articles published by Taiwanese hospital scholars in 2022 and 2023. We evaluated 11,217 articles in 2022 and 9,189 in 2023, all of which were indexed in the Science Citation Index (SCI).
The NPCA method involves comparing network patterns within the IDA framework using similarity coefficients, derived from collaborative maps. These coefficients are then categorized into four levels: identical (values above 0.7), similar (values between 0.5 and 0.7), dissimilar (values between 0.3 and 0.5), and different (values below 0.3). Our analysis revealed that the similarity coefficients for IDA network patterns over the two years were 0.66, 0.83, and 0.28, respectively. This indicates the presence of identical patterns (greater than 0.70) in departments, similarity (above 0.5) in institutes, and distinct patterns (less than 0.3) among authors, particularly when considering first and corresponding contributors, between 2022 and 2023.
The NPCA proves effective in dissecting bibliometric network patterns, underscoring stable trends in departments and institutes, while revealing variations among authors across these years. This methodological advancement opens new avenues for future bibliometric analyses, potentially extending beyond IDA entities to include aspects like team resource management (TRM) in the hospital context.