Semantic Mapping Relational to a Directed Property Hypergraph Model

The real world is mostly a connected data and better be modeled as graph models. Therefore, graph databases are also gaining much attention. To get some advantages of graph models, a huge volume of data in relational models need be converted to graph models. In this paper, we focus on mapping and converting without semantic loss. This approach studies the semantic abstraction of the relational model through its link structure. We summarize that the graph model which is close to the real world is a directed property hypergraph (digraph included). Our approach keeps the semantic abstractions of the real world in the graph model. We conducted some experiments to compare the performances between the naïve graph model approach and a directed property hypergraph. We obtained that the query execution time of directed property hypergraph surpasses the naïve graph model particularly for rather complex queries.