FCA based ontology development for data integration

- A formal and semi-automated method is proposed to support ontology integration.
- The method is designed to deal with data exhibiting implicit and ambiguous information.
- Case studies have been carried out on several non-trivial industrial datasets.
- Resultant ontologies better fit and respect underlying knowledge structure of the domain.

Data is a valuable asset to our society. Effective use of data can enhance productivity of business and create economic benefit to customers. However with data growing at unprecedented rates, organisations are struggling to take full advantage of available data. One main reason for this is that data is usually originated from disparate sources. This can result in data heterogeneity, and prevent data from being digested easily. Among other techniques developed, ontology based approaches is one promising method for overcoming heterogeneity and improving data interoperability. This paper contributes a formal and semi-automated approach for ontology development based on Formal Concept Analysis (FCA), with the aim to integrate data that exhibits implicit and ambiguous information. A case study has been carried out on several non-trivial industrial datasets, and our experimental results demonstrate that proposed method offers an effective mechanism that enables organisations to interrogate and curate heterogeneous data, and to create the knowledge that meets the need of business.

154