Semantic Web in the Age of Big Data: A Perspective: Will We Drown in a Data Tsunami or Enter a Knowledge Utopia?

Abstract

We are awash with "Big Data" to this very day because of the technological advancements made during the past decade. The notion of Big Data 1 refers to the datasets which are gigantic in size to be processed by conventional databases and management techniques (volume), are extremely diverse so that no single data model can capture all elements of the data (variety) and are produced or gathered at an unprecedented scale (velocity) 1. Because of this sheer volume, variety, and velocity of big data, enterprises are facing data heterogeneity, diversity and complexity challenges. However, this big data era came with big opportunities by resolving the associated challenges, so it could transform our traditional way of decision-making. Enterprises with the technical expertise of managing big data are now replacing their usual guesswork and laborious legacy data modeling based decision making processes with facts derived from big data 2. Semantic web technologies 3 such as ontologies help to contextually interpret the heterogeneous big data by associating the data concepts with ontology classes. Ontologies are sets of machine-readable controlled vocabularies that provide the "explicit specification of a conceptualization" of a domain 4. Moreover, ontology orchestrates the domain concepts (both generalized and specialized) in a hierarchical order to foster this hierarchical modeling through the logical relations among them. Such arrangement of concepts provides ontologies with the highest degree of semantic richness of all common models for knowledge representation such as the glossary, topic map, and thesaurus 5. Therefore, semantic mapping (linking of data concepts with ontology classes) not only helps machines to interpret the heterogeneous big data to comprehend the corresponding context but can also help to detect big data anomalies and complete the missing information. Over the past two decades, ontologies have been widely recruited for knowledge representation in various domains ranging from engineering 6-8 , biomedical 9-12 to physics 13,14 and