分類表是圖書館組織圖書資訊的重要工具，然而這類系統多半是以知識的學科分類來建立體系，著重在知識的層屬關係，較少關注使用者角度。本文即嘗試瞭解在分類表的層級結構（Hierarchical Structure）中，發掘讀者認知的分類號關聯（Association）其可能性。初步是以世新大學圖書館所提供兩年大量的借閱記錄為基礎，利用自動化方法分析讀者的相似借閱行為，藉以擷取類號之間的關聯。實驗結果顯示，根據讀者的借閱行為，許多相關類號不僅集中在分類表所設計的類別層級，同時也有不少類號是存在於其它不同的類別層級。針對這些關聯的特性，本文也加以分析歸納，並探討其限制與應用範疇。

Library classification schemes are mostly organized based on disciplines with a hierarchical structure. From the user point of view, some highly related yet non-hierarchical classes may not be easy to perceive in these schemes. This paper is to discover hidden associations between classes by analyzing users’ usage of library collections. The proposed approach employs collaborative filtering techniques to discover associated classes based on the circulation patterns of similar users. Many associated classes scattered across different subject hierarchies could be discovered from the circulation patterns of similar users. The obtained association norms between classes were found to be useful in understanding users' subject preferences for a given class. Classification schemes can, therefore, be made more adaptable to changes of users and the uses of different library collections. There are implications for applications in information organization and retrieval as well. For example, catalogers could refer to the ranked associated classes when they perform multi-classification, and users could also browse the associated classes for related subjects in an enhanced OPAC system. In future research, more empirical studies will be needed to validate the findings, and methods for obtaining user-oriented associations can still be improved.