Knowledge Graph Management Platforms: A Feature Based Review
P.K. Sinha, Sourav Debnath, Kanu Chakraborty, Sagar Bhimrao Gajbe · DLINE eBooks · 2021
Purpose: The purpose of this study is to identify and review the existing knowledge graph (KG) management platforms (KGMP) based on a comprehensive set of features to help KG managers, KG researchers, data scientists, scientific and commercial community of different domain for gauging the platforms for satisfactory coverage and usage.Design/methodology/Approach -The study uses a systematic literature review approach to identify 24 KGMPs and then reviews the platforms, with the help of a set of 32 features prepared from existing literature and by studying the platforms.The set of features were segregated as basic features and core features to make the study granular.Findings-The review revealed that most platforms were developed by corporate companies and very few were available as open-source platforms.The tools have emerged in the last five years, are domain independent, have both on-premise and cloud deployment facilities and are compatible with a variety of operating systems.They have extensive documentations; use cases in various domains with demo versions.The platforms allow creating KG from structured, unstructured and semistructured data.NoSQL databases, especially the standard community accepted graph databases like Neo4J, GraphDB, AllegroGraph, are the preferred choice for data storage.The platforms use standard semantic technologies; for instance, for data representation uses RDF data model; for providing access to data SPARQL endpoint, GraphQL endpoint and for querying data languages like SPARQL, GraphQL.Platforms possess the facilities like metadata management, ontology management, linked data management that are integral for KG management.The platforms also provide KG search and visualization with data analytics facilities by third party integration softwares like KNIME, Tableau etc. Originality/value:This paper carries out a basic level review of KGMPs employing a feature-based approach.Compared to the previous studies, present study not only covers more KGMPs but also provides more features that makes the study granular.