Exploitation of ontological approaches in Big Data: A State of the Art

Abstract

The emergence of web technologies is generating a data deluge called Big Data. All this data is in fact a gold mine to be exploited. However, we are confronted with huge volumes of heterogeneous data (various formats) and varied data (various sources) and in continuous expansion. To deal with this, some research works have introduced ontologies: this is the purpose of this paper. We present the Big Data concept on the one hand and the ontology concept on the other. We first recalled the definitions of Big Data, its main dimensions known by the 3 V (volume, velocity, variety), the fields of application as well as the various problems related to it. We reviewed the different solutions proposed as well as the existing tools by using the NoSQL and the Map-Reduce paradigm implemented in Hadoop and Spark. We then looked at the concept of ontology, starting by recalling the definition of ontology, so an ontology is a conceptual model to represent reality and on which it is possible to develop systems that can be shared and reused. Ontologies are used to represent a domain and reason about its entities. Finally, we presented and discussed some research works that combined ontologies and Big Data. We have found that there is a very abundant literature that deals with big data and ontologies separately, but few studies combine the two concepts together. We will therefore focus on the latter in order to enrich the scientific literature in the domain.