Contextual Domain Knowledge for Incorporation in Data Mining Systems

摘要

The concept of contextual domain knowledge is proposed and incorporated in a generic data mining architecture. Various types of domain knowledge (taxonomies, constraints, previously discovered knowledge and user preferences) are allotted with context information, which is organised in a hierarchical topology. Mediation is defined in order to chose context-dependent domain knowledge for incorporation in a data mining exercise. All components are embedded in a contextual knowledge discovery architecture.