Background (or sometimes referred to as domain) knowledge is extensively used in data mining for data pre-processing and for nugget-oriented data mining tasks: it is essential for constraining the search space and pruning the results. Despite the costs of eliciting background knowledge from domain experts, there has been so far little effort to devise a common exchange standard for its representation. This paper proposes the Background Knowledge Exchange Format (BKEF), a lightweight XML Schema for storing information on features and patterns, and the Background Knowledge Ontology (BKOn), as its semantic abstraction. The purpose of BKOn is to allow reasoning over and integration of analysed data with existing domain ontologies. We show an elicitation interface producing BKEF and discuss the possibilities for integration of such background knowledge with domain ontologies.

Andrej Hazucha @VideoLectures.net