結合自動推薦分類號與主題詞之輔助新書編目系統研究

Abstract

本研究為建立一套輔助編目系統。系統擁有自動推薦分類號與主題詞以及作者欄位的功能。使得整體編目流程縮短時間也增加館員彼此之間編目的一致性。 本文結合了文件檢索的方法，將新書自動賦予了分類號與主題詞，也利用了統計及人工挑選方式自動填入作者欄位。在介面上除了會將推薦結果輸出外，推薦的原因是由於哪本書的支持也會放入。另外再配上相關的分析數據跟外部搜尋引擎結果，讓館員不需要在系統外花額外時間查詢資料。 因此本系統將自動編目的演算法放進來，自動化的單一推薦轉型為各種不同推薦原則，利用資訊檢索的推薦搭配人工挑選，選出最適當的資訊填入。讓研究真正進入圖書編目流程，來驗證此系統的可行性。為了評估系統，本文設計了兩個實驗，並且請館員直接操作系統以評估其效能。

This research is to establish an assistant cataloging system. The system is able to automatically recommend the classification number, subject heading and automatically fill author field. The system can simplify the whole cataloging process and achieve the consistency of librarians' cataloging with each other. This paper apply the method of document retrieval to recommend a new book with one classification number and subject headings, and also apply the statistical and manual selection methods to automatically fill in the author field. In addition to the recommendation results on the interface, the reason for this recommendation is revealed. In addition, the relevant analysis and external search engine results are provided, so that the librarian does not need to reduce extra search steps outside the system. The system uses multiple recommendation ranks instead of single recommendation rank. The librarian can manually select appropriate results from recommendations to fill corresponding field. In order to evaluate the system, two experiments are designed with existing data. Besides, this system is evaluated by librarian.