Seminář: An Ontology for Construction and Reuse of Knowledge Discovery Workflows

Inferring new knowledge from information combined from heterogeneous data sources is one of the key challenges in present-day science. Fusion of such data requires the interplay of diverse specialized algorithms resulting in highly intricate workflows, which are difficult and time-consuming to construct manually. Therefore we propose an ontology of relevant algorithms, knowledge discovery tasks and knowledge/data types to serve as a basis for intelligent computational support of knowledge discovery workflow composition. The ontology can also be used to improve repeatability of experiments and to enable reasoning on the results to facilitate reuse of workflows and results. It is implemented in OWL DL and the structure of workflows is described using OWL-S. We have used the ontology for annotation and planning of knowledge discovery workflows in bioinformatics, product engineering and within Orange data mining toolkit.