Дослідження методів і технологій інтеграції онтологічної моделі з реляційними даними

Abstract

A study of methods and technological solutions on Data Mapping concerning the integration of an ontological and relational data models has been carried out.The main objective of the investigation is to accelerate and reduce the cost of constructing some ontological models for systems that process distributed data in a heterogeneous environment.At present, the theoretical basis for the integration of various data models (Data Mapping) is a di-rection that has been defined as Ontology-based data integration. The theoretical and practical develop-ment of this area is the Ontology-Based Data Access (OBDA) approach, which integrates ontological models presented in the form of RDF graphs with relational data.The methodology of Data Mapping application for distributed data processing systems has been developed. As an example the process of data models consolidation for Ukrainian State Budget monitor-ing system is given. The database of Ukrainian State Budget monitoring system consists of many rela-tional tables, which contain reports of the State Treasury on the implementation of revenue and expendi-ture parts of both state and municipal budgets, as well as the regional section. In addition, the database contains data on lending and arrears of budget institutions.The ontology model connects to data sources through a declarative specification provided in terms of display, including classes and their properties.The given application converts a SPARQL queries into a SQL query to the relational database. The generated SQL query may be execute by the Oracle 11g database driver, which returns the result as a data snapshot. Then, to improve the performance of SPARQL queries, it should be used the semantic query optimization method.Indicative application of the methodology on the example of the construction of an ontological model of the system of monitoring the state budget of Ukraine in Protege 5. It has been demonstrated the results of the execution of the SPARQL-queries to the relational data of the budget process under the Oracle 11g Database. It has been shown the directions semantic optimization for SPARQL-queries, which allow improving obtained data quality.The proposed methodology allows: to integrate data presented by different models — ontological and relational knowledge acquisition; to overcome constraints when databases based on outdated data models are being merged with modern ontological-oriented systems; to eliminate data redundancy in designing knowledge-based systems.