Ontology and trust based data warehouse in new generation of business intelligence: State-of-the-art, challenges, and opportunities

Abstract

© 2015 IEEE. Business intelligence applications are more focused on structured data and support decision makers by providing meaningful information from extracted data mainly coming from day-to-day operational databases and structured external data sources. However, the volume of unstructured data is growing very fast and data analysts need to consider this kind of data especially when analysing external data such as customers' reviews and posts in social media networks and web blogs. As external data is derived from a variety of sources, it is essential to determine the reputation of the source and provide flexibility to the analysts so that they can take into account the trust value of each source in their analysis. Another important consideration is the semantics of extracted textual data from which meaningful information is derived. Ontology is utilized in order to obtain the semantics of textual data. We do not intend to develop ontology from scratch, but rather use the existing available ontologies which are collected in an ontology repository. By using ontology, entities in the extracted textual data are linked with corresponding concepts. Useful knowledge can be inferred and used in the analysis process. In summary, we propose to use the notion of trust to evaluate data sources, and use ontology to enrich textual data. The trusted external data covered global environment, the Voice of the Market, and the Voice of the Customer can be collected and store it in the current data warehouse.