Knowledge discovery for ontology construction

We can observe that the focus of modern information systems is moving from “dataprocessing” towards “concept-processing”, meaning that the basic unit of processing is less and less an atomic piece of data and is becoming more a semantic concept which caries an interpretation and exists in a context with other concepts. As mentioned in the previous chapter, an ontology is a structure capturing semantic knowledge about a certain domain by describing relevant concepts and relations between them. Knowledge Discovery (KD) is a research area developing techniques that enable computers to discover novel and interesting information from raw data. Usually the initial output from KD is further refined via an iterative process with a human in the loop in order to get knowledge out of the data. With the development of methods for semiautomatic processing of complex data it is becoming possible to extract hidden and useful pieces of knowledge which can be further used for different purpose including semiautomatic ontology construction. As ontologies are taking a significant role in the Semantic Web, we address the problem of semi-automatic ontology construction supported by Knowledge Discovery. This chapter presents several approaches from Knowledge Discovery that we envision as useful for the Semantic Web and in particular for semi-automatic ontology construction. In that light, we propose to decompose the semi-automatic ontology construction process into several phases. Several scenarios of the ontology learning phase are identified based on different assumptions regarding the provided input data. We outline some ideas how the …