Benchmarking applied to semantic conceptual models of linked financial data

Abstract

Semantic modeling plays a central role in knowledge-based systems where information sharing and integration is a primary objective. Ontology and metadata description languages such as OWL (Web Ontology Language) and RDF(S) (Resource Description Framework Schema) are commonly the most used for representing semantic models and data. The graph-like structure adopted for semantic metadata representation allows simple and expressive queries by using SPARQL-based subgraph matching. While performance of such knowledgebased systems depends on multiple factors, in this work we present a mechanism to properly choice a semantic modeling pattern in order to significantly reduce the data query execution time. Based on this understanding, this work proposes a comparative analysis of different conceptual modeling approaches on the basis of financial domain. In order to show the efficiency/accuracy of our approach, an evaluation of SPARQL-based queries was performed against different modeled datasets.