Connecting Knowledge to Data Through Transformations in KnowID: System Description
- PDF / 1,978,201 Bytes
- 7 Pages / 595.276 x 790.866 pts Page_size
- 4 Downloads / 212 Views
SYSTEMS DESCRIPTION
Connecting Knowledge to Data Through Transformations in KnowID: System Description Pablo R. Fillottrani1,2 · Stephan Jamieson3 · C. Maria Keet3 Received: 31 December 2019 / Accepted: 11 June 2020 © Gesellschaft für Informatik e.V. and Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract Intelligent information systems deploy applied ontologies or logic-based conceptual data models for effective and efficient data management and to assist with decision-making. A core deliberation in the design of such systems, is how to link the knowledge to the data. We recently designed a novel knowledge-to-data architecture (KnowID) which aims to solve this critical step through a set of transformation rules rather than a mapping layer, which operate between models represented in EER notation and an enhanced relational model called the ARM. This system description zooms in on the novel tool for the core component of the transformation from the Artificial Intelligence-oriented modelling to the relational database-oriented data management. It provides an overview of the requirements, design, and implementation of the modular transformations module that straightforwardly permits extension with other components of the modular KnowID architecture. Keywords Ontology-mediated data access · Data management · Conceptual modeling
1 Introduction Application ontologies, or logic-based conceptual data models, as well as knowledge graphs, are used to achieve effective data management with faster data analysis thanks to querying at the conceptual layer. This advantage comes at a cost of devising a good and efficient way to connect the knowledge to the data; an introductory overview of several options and considerations are described in [12]. The very recently proposed KnowID architecture [9] (see Fig. 1) is a highly modular architecture for such applications, which takes Extended Entity Relationship (EER) diagrams as application ontologies and connects them to the data layer via the so-called ‘Abstract Relational Model’ (ARM) of [1] that uses special object identifiers and a strict extension to * C. Maria Keet [email protected] 1
Departamento de Ciencias e Ingeniería de la Computación, Universidad Nacional del Sur, Bahía Blanca, Argentina and Comisión de Investigaciones Científicas, Bahía Blanca, Argentina
2
Comisión de Investigaciones Científicas, Buenos Aires, Provincia de Buenos Aires, Argentina
3
Department of Computer Science, University of Cape Town, Cape Town, South Africa
SQL for path queries (SQLP) that simplify querying [14]. Set within this context of the KnowID architecture, the main contribution of this system description paper is the presentation of the first proof-of-concept implementation of connecting the AI-oriented knowledge layer to the database-oriented data layer through transformations, transforming application ontologies represented in EER to ARM and vice versa. It shows that what ought to work theoretically, indeed does so practically. It has a front-end for model
Data Loading...
