id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
cord-033721-o1c7m9wy	Kostovska, Ana	Semantic Description of Data Mining Datasets: An Ontology-Based Annotation Schema	2020-09-19		.txt	text/plain	4482	253	49	To semantically describe a DM dataset, we consider three different types of vocabularies/ontologies: (1) vocabularies for annotation of provenance information, such as title, description, license, and format; (2) ontologies for annotation of datasets with DM-specific characteristics, i.e., data mining task, datatypes, and dataset specification; and (3) ontologies for annotation of domain-specific knowledge that helps to contextualize the data originating from a given domain. After describing the four characteristics that govern the modeling of the taxonomies of datatypes, data specification, and tasks, we provide an illustrative example that shows how we can combine them in a single annotation schema for the purpose of semantic annotation of DM datasets. To represent the MTR task and MTR dataset specification, we use the classes defined in OntoDM-core, and connect them with the corresponding datatype class from OntoDT (in our case OntoDT: feature-based completely labeled data with record of numeric ordered primitive output) (see Fig. 7 b) .	./cache/cord-033721-o1c7m9wy.txt	./txt/cord-033721-o1c7m9wy.txt