Refactoring Methods for Knowledge Bases

Abstract

Abstract. The manual development of large knowledge systems is a difficult and error-prone task. In order to facilitate extensions to an existing knowledge base the structural design of the implemented knowledge needs to be improved from time to time. However, experts are often deterred even from important design improvements since some restructurings are too complex to handle. In this paper, we introduce a framework that allows for automated refactorings. Refactoring methods are well-defined and are executed in a semi-automated way. In this manner, the developer is supported during the process of restructuring of even large knowledge bases. Refactoring methods are usually applied to improve the design of the knowledge base; in this paper, we sketch some design anomalies that identify poor design of the knowledge base.

Engineering intelligent systems on the knowledge formalization continuum

“…• Refactoring methods to support manual changes explicit knowledge without changing the intended semantics (e.g., Gil and Tallis, 1997;Baumeister et al, 2004;. They are often used to accomplish vertical transitions to a semantically equivalent version within the same knowledge representation, but are also helpful to restructure knowledge to a less/more formalized level.…”

Empirical Knowledge Engineering: Cognitive Aspects in the Development of Constraint-Based Recommenders

“…Knowledge base refactoring approaches [1] follow the goal of improving the structural design of given knowledge base. Due to repeated maintenance activities, the overall quality of the knowledge base deteriorates and mechanisms are needed that help to restore quality through a set of reorganization steps.…”

“…Due to repeated maintenance activities, the overall quality of the knowledge base deteriorates and mechanisms are needed that help to restore quality through a set of reorganization steps. [1] discuss knowledge base design anomalies and provide detailed recommendations on how to deal with such situations. Examples for such anomalies are objects never used by the application, attribute domain values nearly never used, etc.…”

“…We are considering scenarios where knowledge engineers are interacting with a knowledge acquisition environment and try to develop and maintain recommender knowledge bases. 1 An effect occurring in this context is the so-called knowledge acquisition bottleneck which denotes the communication overhead between domain experts and knowledge engineers in the phase of implementing and adapting knowledge bases. This is based on the fact that domain experts know about recommendation rules and heuristics, but it is very hard for them to translate those into a corresponding executable representation, but then knowledge engineers have a limited knowledge about the product domain.…”

Verification and Refactoring of Ontologies with Rules

“…[11]. Recently, a first step was taken to transfer this idea to the conceptual properties of rule-based knowledge [13] and OWL ontologies [14], respectively.…”