Murat Osman Unalir

In corporate organizations, the integration of customer data stored across distributed systems is critically important for analysis and effective decision making processes. This study proposes a graph-based solution capable of deduplicating customer data collected from various sources in a flexible way. The developed architecture is built upon a graph data model where customer entities are represented along with all their attributes. Matching rules are designed to be dynamically definable and are supported by both exact and similarity based matching strategies. The proposed structure is easily understandable and adaptable by both technical teams and business units. Incomplete, incorrect or structurally inconsistent data from different sources are analyzed through flexible rule combinations to form customer clusters representing the same individual. To preserve raw data and ensure efficient resource usage during the clustering phase, a unified node and relationship type is specified for the matching step. The proposed system provides significant benefits such as improving data quality, identifying duplicated records and ensuring legal compliance. At the same time, it forms a strong foundation for ontology based or large language model supported matching approaches.

Ontology Based Data Access (OBDA) is the provision of data access and data integration as a result of the mapping that is established between an ontology and a data source. Thus, storing large amounts of data becomes easier, more powerful queries can be written, and management of complex information systems can be performed quickly and effectively by using Semantic Web technologies. Ontology Based Access Control (OBAC) uses Semantic Web technologies to enable the enforcement of access control mechanism. Therefore, only authorized persons can access data to protect data privacy. In this study, OBDA and OBAC are integrated to improve security while providing data virtualization with a data model-independent access control approach. Therefore, a use case study for the healthcare domain is presented. Hence, a relational database for the hospital domain, a Hospital Ontology for the related hospital database and an access control policy are created. Also, the relevant mappings between the hospital database and the Hospital Ontology are established by using the Ontop framework and finally, various queries are executed by using Ontop SPARQL to evaluate mappings and access rules.

With the considerable growth of linked data, researchers have focused on how to increase the availability of semantic web technologies to provide practical usages for real life systems. Question answering systems are an example of real-life systems that communicate directly with end users, understand user intention and generate answers. End users do not care about the structural query language or the vocabulary of the knowledge base where the point of a problem arises. In this study, a question answering framework that converts Turkish natural language input into SPARQL queries in the geographical domain is proposed. Additionally, a novel Turkish ontology, which covers a 10th grade geography lesson named Spatial Synthesis Turkey, has been developed to be used as a linked data provider. Moreover, a gap in the literature on Turkish question answering systems, which utilizes linked data in the geographical domain, is addressed. A hybrid system architecture that combines natural language processing techniques with linked data technologies to generate answers is also proposed. Further related research areas are suggested.

Background: One of the state-of-the-art techniques for question answering frameworks is using linked data by converting the user input into SPARQL which is the query language for linked data. Objective: The main target is to emphasize the most fundamental issues while developing a question answering frameworks that accept input in natural language and converting it into SPARQL. Methods: The trend of applying linked data as a data source is gaining popularity among the researchers. In this study, question answering frameworks that combine both natural language processing techniques and linked data technologies are examined. Common principles of examined question answering frameworks recognize user intention, enriching natural language input and converting it to a SPARQL query. Results: 9 studies are selected for further examination to be compared by using selection criteria defined in the research methodology. Conclusion: Resulting outcomes are represented and compared in detail. In addition to the comparative review of systems, a general architecture of question answering frameworks on the linked data is drawn as an outcome of this study to provide a guideline for the researchers who are studying related research fields.

This paper proposes a novel method for assessing text coherence. Central to this approach is an ontology‐based representation of text, which captures the level of relatedness between consecutive sentences via ontologies. Our method encompasses annotating text using ontological concepts and assessing text coherence based on relatedness measurement among these concepts. The ontology‐based relatedness measurement method used in this study considers various types of relationships in ontologies and derived relationships via an inference engine for computing relatedness. We hypothesized that rich variety of relationships and inferred facts in ontologies would improve the success of text coherence assessment. Our results demonstrate that the use of ontologies yields to coherence values that have a higher correlation with human ratings.

Semantic web aims that web contents can be understood and processed by the other software, are to be interpreted, usable and sharable. Security of information in semantic web is normally provided by access control mechanism. OBAC (Ontology Based Access Control) is a model for accessing ontologies. RBAC (Role Based Access Control) is a model for accessing relational databases. In this study, HAC (Hybrid Access Control) is aimed to be applied for mapping of relational database and ontology by means of both OBAC and RBAC together. Access control of roles and profiles are aimed to be managed in HR (Human Resources) domain. To achieve the identified goal, a method is proposed and the tool software is developed within the scope of this study. HAC is improved as an access control by using both RBAC and OBAC. A software architecture is presented for the access of framework. Roles are managed in HR domain by using hybrid data and a solution for database management is proposed.

Today’s technological improvements have made ubiquitous healthcare systems that converge into smart healthcare applications in order to solve patients’ problems, to communicate effectively with patients, and to improve healthcare service quality. The first step of building a smart healthcare information system is representing the healthcare data as connected, reachable, and sharable. In order to achieve this representation, ontologies are used to describe the healthcare data. Combining ontological healthcare data with the used and obtained data can be maintained by storing the entire health domain data inside big data stores that support both relational and graph-based ontological data. There are several big data stores and different types of big data sets in the healthcare domain. The goal of this paper is to determine the most applicable ontology data store for storing the big healthcare data. For this purpose, AllegroGraph and Oracle 12c data stores are compared based on their infrastructural capacity, loading time, and query response times. Hence, healthcare ontologies (GENE Ontology, Gene Expression Ontology (GEXO), Regulation of Transcription Ontology (RETO), Regulation of Gene Expression Ontology (REXO)) are used to measure the ontology loading time. Thereafter, various queries are constructed and executed for GENE ontology in order to measure the capacity and query response times for the performance comparison between AllegroGraph and Oracle 12c triple stores.