The structure of a knowledge base for cataloging rules

作者:

Highlights:

摘要

This article reports on a pilot study that investigates the general patterns and the level of applicability of rules in the AACR2. It studies those rules in the AACR2 related to the title and statement of responsibility area. It attempts to answer two research questions: (1) What is a rule in the AACR2? (2) What are the application domains for the rules? Rule numbers in the AACR2 are divided into two kinds: the rule numbers followed by text and those followed by a set of subrules. Prototype expert systems designed for cataloging as reported in literature have used the former as the defining unit in their knowledge base. However, the study finds that a rule number with text usually consists of many pairs of condition and action (C/A); each pair is a subrule. The analysis of actual application of rules further suggests that the level of applicability of a rule is usually limited. Rules are designed for five major application domains: definition, description, organization, source of information identification, and transcription. Findings of this study suggest that a knowledge base for cataloging rules must base its defining unit on the C/A pair and not the rule number. The structure of each C/A pair in the knowledge base must contain the application domain, the procedural interpretation, the level of applicability, and the condition and action as stated in the text.

论文关键词:

论文评审过程:Received 29 March 1990, Accepted 18 July 1990, Available online 18 July 2002.

论文官网地址:https://doi.org/10.1016/0306-4573(91)90033-I