Creating Data-Driven Ontologies An Agriculture Use Case
Combining various, heterogeneous data sources is a challenge, because data from different sources usually have different syntax and semantics. This also holds for the horticultural domain in which data from various sources and organizations needs to be combined, including greenhouse climate, outside weather, labor usage, plant growth, yield, production, etcetera.
An innovative and promising manner of expressing semantics makes use of an ontology in which concepts and their relations can be modelled in a formal way. When semantics of data from different sources can be aligned, big data analysis is made possible in an unambiguous way. Developing an ontology is usually done using human domain experts that model their knowledge of the domain in an ontology. In order to support this process, we developed an automatic, data-driven, ontology generation toolset that extracts an ontology from a set of domain documents. This toolset makes use of natural language processing text mining algorithms to generate a common horticultural ontology to be used by the sector.
Please find the full article in the output library.