Problem: Usually organizations look at key performance metrics about customers, partners and various internal divisions – but are unable to capture data-driven insights into what is driving these metrics. Moreover, these tools do not capture knowledge on what has worked historically and what has not worked. Obtaining insights (or prescriptive analysis) that are actionable is a valuable product for an organization rather than data or knowledge.
Key Idea: For generating better insights domain knowledge plays a key role. Ontology is a Semantic Web technology that captures domain knowledge more formally. So, using ontologies for generating insights will be more beneficial to the industry.
To achieve the proposed key idea we are working on the following:
Construction of ontology for a given context (e.g. sports, medical, wireless networks, university ranking, and smart buildings) and load the ontology with individuals (i.e., instances.
Based on the context domain– an insight rule (aggregated information) generation algorithm will be developed and applied on the RDF triples, to extract insights at the level of aggregate individuals. Simultaneously, these insights about the individuals is used to update the ontology.
Develop a proof-of- concept for insight-as- a-service where the user can upload their structured data and get valuable insights.