dc.description.abstract | Knowledge Graphs (KG) have been widely popular for presenting real-world entities as nodes and edges according to semantic web rules and regulations. Many organizations and industries have used Knowledge Graphs (KGs) for publishing their datasets according to Linked Open Data (LOD) principles. Many Graph data models, e.g., RDF, Property Graph, and Domain Graph data model, have been used to model real-world entities as Knowledge Graphs (KGs). However, each data model has represented the knowledge differently, which sometimes affects the performance of the Knowledge Graph, especially in data storage and retrieval. The selection of an exemplary graph data model for representing Knowledge Graphs (KGs) plays a vital role in extracting and integrating data from various sources. Wikidata (Vrandečić & Krötzsch, 2014) is a Knowledge Graph representing real-world entities and connecting them to Wikipedia articles. Wikidata entities are defined by the Pages, describing the information as statements. Each statement has some additional information, e.g., qualifiers and references. Wikidata is one of the most extensive Knowledge Graphs where the data is updated daily. Hence, the representation of Wikidata entities in the different graph data models is challenging and costly in terms of data storage and data retrieval. So, the thesis represents the Wikidata in three graph data models, e.g., RDF, Property Graph, and Domain Graph, and does a qualitative analysis of three graph data models by conducting comparison and describe their advantages and disadvantages. The RDF data model represents Wikidata as triples (subject-predicate-object) and uses RDF reification to model Wikidata complex statements. The property Graph data model uses node and edge labels to represent Wikidata entities and model the complex statement as edge attributes and a compact data model. The Domain Graph data model uses the edges as nodes to model Wikidata statements. The RDF reification lacks internal structure and generates many redundant triples, which increases the data storage and reduces the query response time. The Property Graph data model needs to be fully represented Wikidata statements as edge attributes. However, the Domain Graph data model facilitates the edges as nodes, fully represents Wikidata statements, and provides better storage and query response time than RDF and PG. In addition, the thesis represents the general qualitative analysis between three graph data models (RDF, Property Graph, and Domain Graph), which helps the readers to select the best graph data model for modeling Knowledge Graphs (KGs). | eng |