以關聯規則為基發掘圖書館讀者適性書籍

Finding Library Readers' Adaptive Books Based on Association Rules

讀者和服務人員之間互動是圖書館經營理念的核心價值，而館藏資訊的傳播扮演提升圖書館服務品質的輔助角色，在不同讀者需求資訊的差異性下，善用個人化服務技術調整資訊需求是目前圖書館重要提供功能之一。本研究以圖書館讀者之借閱資料為探勘資料來源，以某一讀者為探勘目標，利用資料探勘中的關聯規則（association rule）分別從以下兩方面發掘讀者適性書籍：一是考量書籍是否出現在此讀者的借閱資料中，文中設計一個方法探勘關聯規則，其前置項目集必須被包含於此讀者的借閱資料中，根據關聯規則顯示出的傾向特徵可發掘此讀者適性書籍；二是增加考量讀者曾經借閱書籍的次序性，文中設計一個方法探勘次序關聯規則，其前置項目集必須被包含於此讀者的借閱資料中，根據次序關聯規則顯示出的傾向特徵可發掘讀者適性次序書籍。文中根據提出的方法，以南部某一科技大學圖書館的借閱資料為例，設計與建置一個發掘讀者適性書籍系統，並評估探勘結果的效益。本研究探勘結果，對圖書館規畫讀者個人適性書籍，進而以主動推薦訊息方式吸引讀者借閱，可以提供非常有用的參考資訊。

The core value of the library management is the interaction between readers and service personnel. The dissemination of collection information plays a supporting role to improve the quality of library services. Under the different information of different readers' needs, using of personalized service technology to adjust information services is one of the important functions of libraries. In this paper, readers' borrowing history records are used as the source data of mining. We let a reader as the target of mining, and use association rules to find the reader's adaptive books from two aspects. One is to consider whether books are contained in borrowing history records or not. A method is designed to mine association rules whose antecedents have to be contained in the reader's borrowing history record. According to the characteristics of the association rules, we can find the adaptive books for the reader. The other is to consider extra the sequences of books in borrowing history records. A method is designed to mine sequences association rules whose antecedents have to be contained in the reader's borrowing history record. According to the characteristics of the sequence association rules, we can find the adaptive sequences books for the reader. According to the proposed method, we take borrowing history records of a university library as an example to design and build a system for finding library readers' adaptive books, and evaluate the benefits of the mining results. The mining results of this paper to plan the library reader's personal adaptive books, and then use the active recommendation message to attract readers to borrow books can provide very useful reference information.