Authors:
Alexander Paulus
;
André Pomp
;
Lucian Poth
;
Johannes Lipp
and
Tobias Meisen
Affiliation:
RWTH Aachen University, Germany
Keyword(s):
Semantic Computing, Semantic Model, Knowledge Graph, Ontologies, Semantic Similarity.
Related
Ontology
Subjects/Areas/Topics:
Cloud Computing
;
Coupling and Integrating Heterogeneous Data Sources
;
Data Communication Networking
;
Databases and Information Systems Integration
;
Enterprise Information Systems
;
Internet of Things
;
Semantic Web Technologies
;
Sensor Networks
;
Services Science
;
Software Agents and Internet Computing
;
Software and Architectures
;
Telecommunications
Abstract:
In the context of the Industrial Internet of Things, annotating data sets with semantic models enables the automatic interpretability and processing of data values and their context. However, finding meaningful semantic concepts for data attributes cannot be done fully automated as background information, such as expert knowledge, is often required. In this paper, we propose a novel modular recommendation framework for semantic concepts. To identify the best fitting concepts for a given set of labels, our approach queries, weights and aggregates the results of arbitrary pluggable knowledge bases. The framework design is based on an intensive review of labels that were used in real-world data sets. We evaluate our current approach regarding correctness and speed as well as stating the problems we found.