本論文以讀者之借閱資料為探勘資料來源,每一筆借閱資料包含讀者曾經借閱的書籍,利用興趣加權探勘技術分別從以下兩方面發掘書籍借閱的適性化推薦:一是以某一讀者為探勘的目標,本研究提出一個探勘興趣加權關聯規則(association rules with weighted interest-items)的演算法,興趣加權關聯規則之前置項目組必須包含於此讀者的借閱資料中,根據此興趣加權關聯規則所顯示出的傾向特徵,可發掘此讀者之書籍借閱的適性化推薦;二是以某一書籍為探勘目標,本研究提出一個探勘興趣加權關聯規則的演算法,興趣加權關聯規則之後置項目組必須為此書籍,根據此興趣加權關聯規則所顯示出的傾向特徵,可發掘適性化借閱此書籍的讀者。本研究根據所提出的兩個探勘方法,設計與建置一個發掘書籍借閱的適性化推薦系統。此探勘結果,對圖書館擬定書籍借閱的適性化推薦,將可以提供非常有用的參考資訊。
In this paper, readers' borrowing history records are used as the source data of mining. Each borrowing history record contains a reader once borrowed books. The study uses mining techniques with weighted interest-items to find adaptive recommendations of borrowing books from two aspects. One is to make one reader as the target of mining. The study presents a method to mine association rules with weighted interest-items whose antecedents are contained in the reader's borrowing history record. According to the characteristics of the association rules with weighted interest-items, the adaptive recommendations of borrowing books can be found for the reader. The other is to make one book as the target of mining. The study presents a method to mine association rules with weighted interest-items whose consequents are the book. According to the characteristics of the association rules with weighted interest-items, the adaptive readers can be found for borrowing the book. A mining system is designed and constructed for finding adaptive recommendations of borrowing books according to the both methods. The results of the mining can provide very useful information to plan the adaptive recommendations of borrowing books for libraries.