With the increase in data, knowledge management and system integration become challenging. Although the use of Virtual Knowledge Graph, via Ontop, is a promising solution, its application is complex. The article introduces ApplyOnTop, an approach for semantic integration of databases based on Ontop. This approach proposes the execution of structured activities that assist in applying Ontop. To validate the applicability of ApplyOnTop, an experiment was conducted, using the approach in a real-world database integration scenario. The results demonstrate improvements in ontology creation and linking with databases, systematizing the use of Ontop for more efective integration.
For this work, 20 studies adhering to its theme were selected. These studies underwent an evaluation stage, which included the application of six classification criteria listed below.
1. Did the work use a relational database as a data source? 2. Did the work generate an ontology? 3. Did the work present a detailed process for ontology construction? 4. Did the work link the defined ontology to the database? 5. Did the work enable querying the database from the semantic layer defined by the ontology? 6. Did the work define a detailed process for linking the ontology to the database?
Among the analyzed studies, all incorporated the use of ontologies. Two used existing ontologies, while the others developed their own. All allowed querying of data through the generated semantic layer, with most using a database. However, none detailed the process of ontology construction and its linkage to the database.
The ApplyOnTop is an approach designed to streamline database integration through the use of Ontop. It guides users through a set of structured steps aimed at making the mapping process between ontology elements and data stored in databases more accessible.
ApplyOnTop is built with a focus on domains with multiple developed information systems. Its strategy involves constructing the ontology from macro views of the individual ontologies of each system involved in the domain. Subsequently, these macro views of the systems’ ontologies are merged and detailed to generate the domain ontology, which is then linked to the systems’ databases.
ApplyOnTop is divided into two distinct phases, as illustrated in the Figure 1. Phase 1, called Domain Ontology Construction, encompasses activities related to obtaining domain knowledge and elaborating its representation in the form of an ontology. Phase 2, Ontology-Database Linkage, includes activities involved in mapping the ontology and linking it to the data’s physical structure.
ApplyOnTop allows for evaluation, refinement, and continuous improvement of products developed in previous stages, flexibilizing their sequence. As domain knowledge progresses, there is a need to revisit and improve products, especially the ontology, initially constructed with a business perspective and later enhanced with the technical perspective of databases.
The ontology construction process is segmented into four distinct phases: Specification,
Conceptualization, Formalization, and Implementation. In the Specification phase, the domain scope is delineated,
the ontology’s purpose is defined, and knowledge sources are identified [
The ApplyOnTop approach aims to unify the ontologies of diferent systems involved into a single representation, thus simplifying the integration process. After each stage, a review and refinement are conducted to ensure that the ontology meets the established requirements. This approach allows for continuous evolution of the ontology as domain knowledge deepens. At the end of the process, users have a robust and functional ontology ready to facilitate data integration and querying eficiently and efectively.
This phase was inspired by Methontology [
This phase aims to associate ontology elements with data stored in domain databases, facilitating data querying and understanding. Linkage is essential for the Ontology-Based Approach and requires creating mappings between ontology elements and tables. This phase consists of four stages: Data Identification and Preparation, Data Mapping, Query Development, and Verification.
In the Data Identification and Preparation stage, data sources are made available and analyzed. In
this stage, the Physical Data Model Diagram is used to aid in understanding the data [
The Query Development stage aims to formulate queries capable of retrieving information from the databases and eficiently answering questions defined for the ontology. SQL queries compatible with the database structure and SPARQL queries to interact with the ontology and associated data are elaborated. Finally, in the Verification stage, tests are conducted to verify the accuracy of the linkage between the ontology and the databases. Throughout the process, opportunities may arise for ontology evolution to a more robust and domain-appropriate representation.
For the experiment, the domain of Contracts and Customer Support was chosen. An ontology for this domain, named Ontology for Contracts and Customer Support Integration, was developed as the foundation for the experiment. Following ontology construction, steps were taken to link it to the relevant databases.
ApplyOnTop proved efective in integrating the systems within the domain, as evidenced by the results obtained throughout the proposed phases and stages. During the Domain Ontology Construction phase, a comprehensive view of the entities contained in the systems was obtained. ApplyOnTop introduced defined stages that enhanced domain understanding, initially encompassing business perspective through the use of Conceptual Map and Preliminary Ontology Design. This understanding contributed to the elaboration of the Ontology Design, covering the classes identified in the domain’s systems.
An important aspect of ApplyOnTop was its flexibility and iterative format, allowing for continuous adjustments and improvements. The application of proposed diagrams in ApplyOnTop, such as Ontology Design and Ontology-Database Linkage Diagram, significantly aided in the analysis and construction of both ontology and mappings. This contribution was crucial due to the ease of understanding and visualization of the necessary relationships and linkages for enabling integration.
This outcome validates the developed approach, ofering organizations an option to efectively and accurately address the complex integration demands present in their businesses. ApplyOnTop provided a systematic view of the steps required to enable semantic integration using Ontop.
This work aimed to develop an approach for semantic database integration based on Ontop, named ApplyOnTop. This approach proposes the definition of structured activities to help users minimize the complexities inherent in semantic data integration, making the application of Ontop more systematic.
To validate the functionality of ApplyOnTop, an experiment was conducted, utilizing the need for integrating two systems within a company. The experiment was successful, demonstrating the clear creation of the ontology and its linkage to databases through the use of activities proposed in the approach.
ApplyOnTop presented significant aspects such as the creation of diagrams aiding analysis and decision-making during Ontop integration execution. For example, the Conceptual Map plays a significant role in obtaining initial domain information. Two other diagrams created were the Preliminary Ontology Design and the Ontology Design, allowing views of partial ontologies and the consolidated ontology to be built, aiding in analysis and ontology configuration in Protégé. Additionally, the construction of the Ontology-Database Linkage Diagram assists in mapping ontology elements with stored data.
Ultimately, the significant characteristic of the work was the creation of a systematic approach for Ontop usage in semantic integration. This systematization defines phases and steps to be executed, allowing for an iterative approach that enables revisiting previous steps to enhance already built products. Thus, as knowledge about the domain, business, and database is obtained, the ontology may need to evolve.
The ApplyOnTop approach presented a clear structure and defined guidelines to simplify Ontop usage. By establishing specific steps to follow, the process of ontology construction and linkage becomes more accessible and guided. Therefore, this work can assist professionals interested in working with Ontop, simplifying and facilitating the complex journey of data integration. 5.1. Contributions 1. Detailing for Ontology Construction Using Menthontology: Although Menthontology was adopted as the foundation, the ontology construction process has been enhanced. In this enhancement, the focus is on building ontologies in domains that encompass multiple information systems. To achieve this, steps have been included to guide the ontology builder from documenting the acquired knowledge, using conceptual mapping, to the actual construction of the ontology, supported by Preliminary Ontology Design and Ontology. 2. Visualization Aspect of Ontology Design: The use of Ontology Design facilitates analyzing and planning the ontology to be built in an easier way. This approach allows for a preview visualization of classes and relationships before implementation in Protégé, significantly simplifying the process. Thus, not only does it reduce the efort involved in construction, but it also creates an important input that can be used alongside the created ontology, contributing to a clearer conception. 3. Integration of Partial Ontologies Based on Databases: The strategy of developing macro views of the ontologies from each system involved in the domain, to later merge them and generate the final ontology, aids in the creation process. This approach helps to make the process clearer and more precise, allowing for a comprehensive understanding of the individual ontologies before their consolidation. 4. Link Visualization: The use of the Link Diagram enables the analysis and documentation of the mapping to be established between the ontology and the databases. This tool makes the linking process more transparent, providing a visual representation that facilitates understanding of the relationships between the ontology elements and the data in the corresponding bases.