研究型大學圖書館期刊評選模式之研究

Abstract

研究型大學每年均須面臨期刊增刪的問題，尤其西文期刊價格昂貴，如何有效追蹤與分析所採購的期刊之使用情形等資訊，供決策者參考，為圖書館重要之課題。本研究以引文分析為基礎，提出使用者為導向的期刊評選模式，考慮：一、機構使用者引用數量；二、國內專家所提出核心期刊清單；三、國際上JCR所蒐集的高影響係數（Impact Factor）期刊…等三個面向，最後再以彈性度較高的權重法決定最後的期刊購買清單。 本研究最後以交通大學資工系近五年來的博碩士論文，共27,016筆引用文獻資料匯入系統模式中，得出的分析結果顯示：一、參考文獻中，利用布萊德福分區運算出前二區的62種期刊，僅占全部被引用期刊的4.22%；二、以本研究模式評分結果顯示，交大採購清單符合率約占82.26%，但若納入電子資源收錄之期刊做比對，符合率則高達100%；三、由近五年的引用情形得知，將近有30種資工系所推薦採購的期刊未被引用過，應可作為未來刪減期刊之參考；四、本研究所建置之系統，具備引文資料分析功能且提供適度評分彈性，可做為後續期刊評選決策支援之用。

Every year research-oriented university faces the difficulty of additions and deletions of journals. Considering the high price of English journals, efficiently tracing and analyzing the usage of purchased journals becomes one of the main library tasks. Based on citation analysis, a reader-oriented journal evaluate model is created in this research. Taking 3 factors, the amount of articles cited by the organization users, a core journal list selected by domestic experts, and the High-impact factor journals in JCR into consideration, the final journal purchased list is created through an additional weighting method. The study processed the references of doctoral and Master’s thesis from computer science department, National Chiao Tung University for the past 5 years, a total of 27,016 bibliographic records. The analysis outcome is as follows: First, 62 journals enlisted in the first two categories through Bradford’s Law occupy only 4.22% of all cited journals. Second, according to this research model, the conformation rate of National Chiao Tung University purchasing list is 82.26% and the ratio goes up to 100% when taking the electric journals into consideration. Third, analyzing the data of the past five years, we conclude that nearly 30 of all journals recommended by the computer science department have never been cited. Such data may serve as the negotiation of additions and deletions of the periodical list. Last but not least, the system developed in this research is provided with citation analysis function and flexible scoring system, which can further assist in journal evaluation decision.